Motor vehicle crashes are a leading cause of injury in the Army. Behaviors increasing risk for motor vehicle crashes are also prevalent, but research has not linked these behaviors directly to injury outcomes (e.g., hospitalizations). To evaluate the relationship between behavior and motor vehicle crash injuries, 99, 981 Army personnel who completed Health Risk Appraisal surveys in 1992 were followed for up to 6 years. Cox proportional hazards modeling was used to evaluate speeding, seat belt use, drinking patterns, and demographics. A total of 429 soldiers were hospitalized for motor vehicle injury. Unadjusted analyses revealed that heavy drinking, drinking and driving, speeding, low seat belt use, younger age, minority race/ethnicity, and enlisted rank were significantly associated with motor vehicle injury, but neither smoking nor gender was. Multivariate models showed a significant trend of increasing injury risk with younger ages. Soldiers under age 21 were injured almost five times more often than those over age 40 (HR 4.89, 2.56-9.33). Also associated with risk for hospitalizations were minority race (HR 1.78, 1.46-2.18), heaviest drinkers versus abstainers (HR 1.81, 1.11-2.94), and seat belt use of 50% or less versus 100% (HR 1.40, 1.07-1.85). Although nonsignificant, there was evidence of an age-drinking interaction where the difference in injury risk between those older and those younger than 21 was greatest at low alcohol consumption levels. Modifiable risk factors associated with motor vehicle injuries include heavy drinking and low seat belt use. Programs targeting these behaviors that meet the needs of young and minority soldiers are needed. The high density of young, at-risk soldiers residing in base housing may provide a unique opportunity for a residential intervention program.