Background:There is limited, and inconsistent, data on the prospective association between physical activity and surrogate markers of adiposity in older adults. We aim to determine the prospective association of leisure time physical activity (LTPA) with body mass index (BMI), waist circumference (WC) and the incidence of obesity. Methods:This prospective analysis included 7144 individuals with a mean age of 67 ± 6.2 years, from the PREvención con DIeta MEDiterránea (PREDIMED) study. BMI and WC were measured and LTPA was recorded using the Minnesota Leisure Time Physical Activity Questionnaire. Exposure and outcome variables were calculated as cumulative average of repeated measurements. Results:Total LTPA was inversely associated (P < 0.001) with BMI and WC. The difference in BMI and WC between extreme quintiles of LTPA (Q1-Q5) was 2.1 kg/m2 (95% confidence interval (CI) 1.68; 2.49, P < 0.001) and 4.8 cm (CI 2.28; 7.25, P < 0.001), respectively. Low-intensity LTPA was inversely associated with BMI but not with WC, while moderate/vigorous LTPA showed an inverse relationship with BMI and WC. The hazard of general and abdominal obesity incidence decreased across quintiles of total and moderate/vigorous LTPA (P < 0.001 for both), whereas low-intensity LTPA was inversely associated with the incidence of general obesity (P < 0.001). Conclusion:LTPA was inversely associated with BMI, WC and incidence of general and abdominal obesity. The finding that low-intensity LTPA was inversely related to BMI and the incidence of obesity is of particular importance because this level of physical activity could be a feasible option for many older adults.