Background: To examine the relationship between social capital and depression among community-dwelling older adults in Anhui Province, China.Methods: A cross-sectional study was conducted among older people selected from three cities of Anhui Province, China using a multi-stage stratified cluster random sampling method. Data were collected through questionnaire interviews and information on demographic characteristics, social capital, and depression was collected. The generalized linear model and classification and regression tree model were employed to assess the association between social capital and depression. Results: Totally, 1,810 older people aged ≥ 60 years were included in the final analysis. Overall, all of the social capital dimensions were positively associated with depression: social participation (coefficient: 0.35, 95% CI: 0.22 - 0.48), social support (coefficient:0.18, 95% CI:0.07 - 0.28), social connection (coefficient: 0.76, 95% CI:0.53 - 1.00), trust (coefficient:0.62, 95% CI:0.33 - 0.92), cohesion (coefficient:0.31, 95% CI:0.17 - 0.44) and reciprocity (coefficient:0.30, 95% CI:0.11 - 0.48), which suggested that older people with higher social capital had a smaller chance to develop depression. A complex joint effect of certain social capital dimensions on depression was also observed. The association with depression and the combinative effect of social capital varied among older adults across the cities. Conclusions: Our study suggests that improving social capital could aid in the prevention of depression among older adults.