Objective This study aimed to investigate whether sitting time, as a form of sedentary behaviour, is related to incident depressive symptoms in older people. Methods This study included 3503 participants (mean age 71.7 years, 50.1% female) from the ‘Obu Study of Health Promotion for the Elderly' cohort study. At baseline and then 15 months later, the participants reported their status of depressive symptoms using the 15‐item Geriatric Depression Scale. During the baseline assessment, the participants were also asked about their sedentary behaviour on weekdays over the past 7 days and, from there, categorized into three groups (<240, 240–480, ≥480 min/day). Demographic data and the other health behaviours were also assessed at the baseline. Results Cross‐sectional analysis revealed that 437 participants (12.0%) had depressive symptoms. In a prospective analysis, the logistic regression model revealed that the odds ratio for depressive symptom incidence was higher in participants who, at baseline, spent 480 min or more per day sitting (1.636; 95% confidence interval [CI] 1.015 to 2.636, p = 0.043), and in those who spent 240–480 min (1.605; 95% CI 1.085 to 2.375, p = 0.018) in comparison with those who spent less than 240 min. Conclusions Sedentary behaviour significantly affects the risk of incident depressive symptoms. Further research is needed to develop an intervention strategy to manage depressive symptoms, as the second most common cause of burden of disease among older adults. Copyright © 2016 John Wiley & Sons, Ltd.