To estimate the hypothetical effects of substituting time spent in moderate-to-vigorous physical activity (MVPA) and time spent in sedentary behavior (SB) and their associations with diabetes. A cross-sectional study using exploratory survey methods was performed in Alcobaça city in the state of Bahia, Brazil in the year 2015. A total of 473 older adults (aged ≥60 years) participated in the study. Diabetes mellitus, time MVPA, and SB were assessed in a self-reported manner. The Poisson regression was used to verify the hypothetical effects of the substitution of MVPA with SB on diabetes. The substitution of the time in MVPA with time in SB showed higher prevalence ratios of diabetes. Conversely, the substitution of the time in SB proved to be a protective factor, with risks reduced by between 4% and 19%. The substitution of the time spent in MVPA with the same amount of time spent on SB can lead to an increase in the probability of diabetes, and a longer reallocation time corresponded to a greater risk.