How to Find Military Service and Pension Records for Genealogy

Military records can identify an ancestor’s unit, rank and service dates, but there is no single database that contains every veteran’s complete story. The right starting point depends on the conflict, branch, type of service and date of separation. A volunteer soldier in the Civil War, a Regular Army enlisted person in the 1890s and an Army veteran of World War II each follow a different research path.

The most productive approach is to separate two questions. First, what proves the person’s service? Second, what additional file might name relatives or document family events? A compiled military service record may answer the first question, while a pension application, bounty land file, state record or later Official Military Personnel File may answer the second.

What kinds of military records should a genealogist look for?

The main record types do different jobs. An index can identify a likely soldier and unit, but it is usually a pointer to a fuller record. A service file follows military activity. A pension application documents a benefit claim and may include evidence supplied by a veteran, widow or heir.

The National Archives guide to military genealogy makes the key jurisdictional distinction: federal records through the early twentieth century are generally in Washington, DC, later personnel files are in St. Louis, and state militia records belong with the appropriate state archive. That division is more useful than beginning with a commercial database and hoping the correct collection appears.

Which facts should you gather before searching?

Begin with enough identifying detail to distinguish your ancestor from people with the same name. Write down the full name used during service, spelling variations, approximate birth date and place, residence before enlistment, branch, conflict, service dates, state, unit and any service number you already have.

Family sources often provide the first clues. Check discharge papers, photographs with insignia, letters, obituaries, grave markers, family Bibles, medals and county histories. An obituary may identify a unit or veterans organization, which is one reason newspapers can be a valuable genealogical resource. Record exactly where each clue came from. A family story that an ancestor served is a lead, not proof that you have found the correct person.

For common names, build a short identity profile before ordering anything. Residence, age, spouse, occupation and known relatives can help separate two soldiers with similar names. Search spelling variants because names were copied from handwritten rolls, and learn the older handwriting forms that may affect transcription. The site’s guide to reading old handwriting explains a careful comparison process.

What does a Compiled Military Service Record tell you?

A Compiled Military Service Record, commonly called a CMSR, is usually the first federal file to seek for a volunteer soldier. According to the National Archives description of CMSRs, the War Department created these files from muster and pay rolls, regimental returns, descriptive books, hospital rolls and other records. NARA holds them for volunteer soldiers from the Revolutionary War through the Philippine Insurrection.

A CMSR may give enlistment and discharge dates, rank, periods present or absent, residence at enlistment, physical description, pay, promotion, wounds, hospitalization, capture or death. It can establish the outline of a military career, but it usually does not identify parents or next of kin. It also rarely proves that a soldier personally fought in every battle associated with the regiment.

For Civil War research, the National Park Service Soldiers and Sailors database is a practical index to names, sides, units and companies. Use it to identify possible records, then follow the result to NARA and other evidence. The National Park Service cautions that linking a person to a regiment does not establish presence at a particular battle. A soldier might not yet have joined, might have been absent, or might have been assigned elsewhere.

Why can a pension file be more useful for family history?

Pension files can be genealogically rich because the claim required proof. Veterans, widows and heirs had to connect the service to the applicant and, where relevant, prove a family relationship. The National Archives pension and bounty land guide says these files may include service narratives, birth and marriage records, death certificates, family letters, witness statements and affidavits.

A widow’s claim may state her maiden name, the date and place of marriage, residence, the veteran’s death date and the names of minor children. An heir’s application may document the deaths of both the veteran and widow. Not every file contains every item, and a rejected claim can still hold useful evidence about what the applicant asserted and what the government questioned.

Service and pension files are separate records, so finding one does not mean you have found the other. For Union Civil War soldiers, NARA identifies the CMSR, pension application and unit records as three major sources. Confederate service records are federal holdings, but Confederate pensions were issued by certain states and are held by those state archives. That difference makes the veteran’s postwar residence important.

Bounty land applications deserve a separate search for earlier service. These claims can contain family evidence similar to pension files. They may also connect military research with land records that document migration and family relationships.

How do you research World War I and later service?

For more recent federal service, the usual target is the Official Military Personnel File, or OMPF, held by the National Personnel Records Center. The current NARA request guide explains that these records may include separation documents such as a DD Form 214, personnel records and medical records. Most remain paper based and are not available online.

Access depends on when the veteran left service. NARA currently treats records as archival and open to the public 62 years after separation. More recent nonarchival files are restricted, with broader access for the veteran or defined next of kin. Researchers should confirm the current rule before requesting another person’s file because privacy and custody status change with time.

A useful request gives NARA the veteran’s complete service name, branch, dates of service, date and place of birth, and service number or Social Security number when known and appropriate. For a deceased veteran, next of kin must also provide proof of death. If you know the last unit, place of entry or discharge location, include those details, especially when a fire related reconstruction may be needed.

What did the 1973 National Personnel Records Center fire destroy?

The 1973 fire was devastating, but it did not destroy every military record. The National Archives history of the fire reports that approximately 16 to 18 million personnel files were destroyed. NARA estimates losses of about 80 percent for Army personnel discharged from November 1, 1912 through January 1, 1960, and about 75 percent for Air Force personnel discharged from September 25, 1947 through January 1, 1964 whose names came alphabetically after James E. Hubbard.

There is no complete list of the lost files. A negative response therefore does not prove that a person did not serve. NPRC uses auxiliary sources, including morning reports, pay records, state records, Veterans Administration claims and organizational files, to reconstruct some basic service information. Reconstruction may recover an outline without recreating every document that once appeared in the original file.

This is a good example of why military genealogy depends on record groups rather than one perfect file. When an OMPF is missing, a researcher can look for draft registration, a unit roster, a county discharge, a VA claim, a burial record, an obituary, correspondence or a state bonus application. Each source answers a smaller question, and their agreement can rebuild the service story.

Where do Upper Midwest and state records fit?

State archives and historical societies are essential when the service was in a militia or state organization, when the federal file is incomplete, or when a veteran filed a state benefit application after returning home. NARA explicitly states that it does not hold state militia records. The correct repository may be in the state where the unit was organized, where the veteran lived after service or where a discharge was recorded.

For an Upper Midwest ancestor, check the catalog of the relevant state archive or historical society, then search county recorders, local libraries, newspapers and veterans organizations. A Minnesota resident could have joined a federal volunteer regiment, served in a state unit, moved across the Wisconsin border after the war, and filed papers in a county courthouse. Residence is evidence about where records may be, not proof of where the person served.

Local records can also solve an identity problem. A federal index might list only a common name and regiment. A county discharge, newspaper notice or veterans association roster may add residence, occupation, spouse or comrades who connect that soldier to the family you are researching.

What is the most efficient research sequence?

Use a sequence that saves the expensive or slow requests until you have a strong identity match.

1. Build the ancestor profile from family papers, census records, obituaries and known relatives.
2. Identify the likely conflict, branch, state and unit. Keep spelling variants and possible aliases together.
3. Search a reliable index, then save the collection title, repository, record identifier and exact result.
4. Obtain the underlying service record rather than stopping at the index entry.
5. Search separately for a pension, bounty land application or later OMPF.
6. Check state archives and local repositories for militia, discharge, bonus and veterans organization records.
7. Compare every new fact with the ancestor profile. Resolve conflicts before combining records into one person’s story.
8. Read the unit history only after confirming the individual and service dates. A unit’s presence at a battle is not proof that the ancestor was there.

The result should be a documented conclusion, not a collection of screenshots. Cite the record series, repository, file or certificate number, image or page and access date. If the file is incomplete or reconstructed, say so. Military records can reveal a remarkable amount about a person’s movements and family, but their limits are part of the evidence too.