OBJECTIVES: To examine the association between alcohol intake and problem drinking history and bone mineral density (BMD), falls and fracture risk. DESIGN: Cross-sectional and prospective cohort study. SETTING: Six U.S. clinical centers. PARTICIPANTS: Five thousand nine hundred seventy-four men aged 65 and older. MEASUREMENTS: Alcohol intake and problem drinking histories were ascertained at baseline. Follow-up time was 1 year for falls and a mean of 3.65 years for fractures. RESULTS: Two thousand one hundred twenty-one participants (35.5%) reported limited alcohol intake (<12 drinks/y); 3,156 (52.8%) reported light intake (<14 drinks/wk), and 697 (11.7%) reported moderate to heavy intake (≥14 drinks/wk) in the year before baseline. One thousand one men (16.8%) had ever had problem drinking. In multivariate models, as alcohol intake increased, so did hip and spine BMD (P for trend <.001). Greater alcohol intake was not associated with greater risk for nonspine or hip fractures. Men with light intake, but not moderate to heavy intake, had a lower risk of two or more incident falls (light intake: relative risk (RR)=0.77, 95% confidence interval (CI)=0.65-0.92; moderate to heavy intake: RR=0.83, 95% CI=0.63-1.10) than abstainers. Men with problem drinking had higher femoral neck (+1.3%) and spine BMD (+1.4%), and a higher risk of two or more falls (RR=1.59; 95% CI=1.30-1.94) than those without a history of problem drinking and similar total hip BMD and risk of fracture. CONCLUSION: In older men, recent alcohol intake is associated with higher BMD. Alcohol intake and fracture risk is unclear. Light alcohol intake may decrease the risk of falling, but a history of problem drinking increased fall risk.