Background The use of self‐report measures of physical activity is a serious methodological weakness in many studies of physical activity and depressive symptoms. It is still equivocal whether light physical activity protects older adults from depressive symptoms. Objective This study aimed to explore whether objectively measured light physical activity, independent of sedentary and moderate‐to‐vigorous activity, is associated with a reduced risk of subsequent depressive symptoms in older adults. Methods This was a 2‐year prospective cohort study. A total of 285 community‐dwelling older adults aged 65 years or older were interviewed in 2012. A second wave of assessment was carried out in 2014 involving 274 (96.1%) participants. Time spent in physical activity at different intensities was assessed using triaxial accelerometers. Depressive symptoms were measured using the 15‐item Geriatric Depression Scale. Negative binomial regression models with adjustment for baseline depressive symptoms, accelerometer wear time, socio‐demographic variables, lifestyle behaviors, and chronic disease conditions were conducted. Results Time spent in moderate‐to‐vigorous and light physical activities were both inversely related to depressive symptoms at follow‐up. Sedentary time was associated with an increased risk of subsequent depressive symptoms. When sedentary or moderate‐to‐vigorous activity were included in the multivariable‐adjusted regression models with light physical activity simultaneously, only light physical activity remained significant. Sensitivity analyses for assessing confounding and reverse causation provided further support for the stability of these findings. Conclusion Light physical activity, independent of sedentary and moderate‐to‐vigorous activity, is associated with a reduced risk of subsequent depressive symptoms in later life. Copyright © 2017 John Wiley & Sons, Ltd.