Cognitive impairment is an important contributor to disability. Limited clinical trial evidence exists regarding the impact of physical exercise on cognitive function (CF). We report results of a pilot study to provide estimates of the relative impact of physical activity (PA) on 1-year changes in cognitive outcomes and to characterize relationships between changes in mobility disability and changes in cognition in older adults at increased risk for disability. Sedentary persons (102) at increased risk for disability (aged 70-89 years) were randomized to moderate-intensity PA or health education. Participants were administered the Digit Symbol Substitution Test (DSST), Rey Auditory Verbal Learning Test (RAVLT), modified Stroop test, and Modified Mini-Mental State Examination at baseline and 1 year. Group differences were not significant but improvements in cognitive scores were associated with improvements in physical function. Specifically, the DSST significantly correlated with change in the Short Physical Performance Battery score (r = .38, p = .0002), in chair stand score (r = .26, p = .012), in balance score (r = .21, p = .046), and in 400-m gait speed (r = .15, p = .147). Change recall on the RAVLT and in the Stroop test was also positively correlated with changes in chair stand and balance, respectively. These results provide further support for the benefits of exercise on CF in older adults. An adequately powered clinical trial of PA involving older adults at increased risk for cognitive disability is needed to expand the indications for prescribing exercise for prevention of decline in brain function.