Using data from the Cardiovascular Health Study, we studied the relation between body composition (fat mass and fat-free mass, assessed by bioelectrical impedance) and self-reported, mobility-related disability (difficulty walking or stair climbing) in 2714 women and 2095 men aged 65-100 y. In a cross-sectional analysis at baseline (1989-1990), disability was reported by 26.5% of the women and 16.9% of the men. A positive association was observed between fat mass and disability. The odds ratio for disability in the highest quintile of fat mass was 3.04 (95% CI: 2.18, 4.25) for women and 2.77 (95% CI: 1.82, 4.23) for men compared with those in the lowest quintile. Low fat-free mass was not associated with a higher prevalence of disability. In a longitudinal analysis among persons not reporting disability at baseline, 20.3% of the women and 14.8% of the men reported disability 3 y later. Fat mass at baseline was predictive of disability 3 y later, with odds ratios of 2.83 (95% CI: 1.80, 4.46) for women and 1.72 (95% CI: 1.03, 2.85) for men in the highest quintile of fat. The increased risk was not explained by age, physical activity, chronic disease, or other potential confounders. Low fat-free mass was not predictive of disability. The results showed that high body fatness is an independent predictor of mobility-related disability in older men and women. These findings suggest that high body fatness in old age should be avoided to decrease the risk of disability.