Longitudinal studies of objectively measured physical activity are lacking in older adults. We tested whether objective measures of total daily activity decline more rapidly in older adults. This prospective, observational cohort study included 519 community-dwelling older persons from across metropolitan Chicago participating in the Rush Memory and Aging Project. Repeated total daily activity measures (leisure and non-leisure physical activity) were derived from actigraphic recordings for up to 10 days. Generalized estimating equation models which controlled for demographics measures were employed. At baseline, age was inversely related with the level of total daily activity (estimate, -0.014, S.E. 0.002, p<0.001). During up to 6 years of follow-up, total daily activity declined by about 0.070 × 10(5) activity counts/day/yr (estimate -0.065, S.E. 0.005, p<0.001). Total daily activity declined 3% more rapidly for each additional year of age at baseline (estimate -0.002, S.E. 0.001, p=0.027). Thus, total daily activity declined almost twice as fast in an individual 91 years old at baseline versus an individual 71 years old. A higher level of education was associated with a slower rate of decline (estimate 0.004, S.E. 0.002, p<0.018). The associations of age and education with the rate of declining total daily activity were unchanged when controlling for baseline level of motor and cognitive function, other late-life activities and chronic health conditions. These data suggest that total daily activity in very old adults declines more rapidly with increasing age. Thus, physical inactivity is likely to become a larger problem in our aging population.