Exploring the social factors of mental health among older adults has become a hot topic. This study aimed to examine the relationships between internet use, social capital and depressive symptoms in older adults. Our data were derived from a sample of 6,840 respondents aged 60 and over in the 2018 wave of the China Family Panel Studies. The ordinary least square (OLS) regression results showed that both Internet use characteristics (including access, emotional activities, and online time) and social capital components (including contact with adult children and trust) were protective factors for the prevention of depressive symptoms among older adults. The generalized structural equation modeling (GSEM) results displayed that Internet use not only had a negatively direct effect on depressive symptoms, but also generated a negatively indirect effect on depressive symptoms by structural social capital (i.e., contact with adult children), suggesting that structural social capital mediated the above link. Conversely, the indirect effects of internet use on depressive symptoms via cognitive social capital (i.e., interpersonal trust and institutional trust) were significantly positive, indicating that the relationship between Internet use and depressive symptoms was suppressed by cognitive social capital. These findings address the gaps in previous research on older adults' mental health and have practical implications for policy makers.