Social participation is associated with cognitive function; however, their causal relationships have not been reported yet. This study was designed to examine the autoregressive effects and bidirectional causal relationship between social participation and cognitive function. In this secondary longitudinal data analysis, we enrolled 4,834 Korean adults. A cross-lagged panel model with fixed effects was used to examine the causal relationships between social participation and cognitive function. Both participation (unstandardized coefficient = .370, p < .001) and cognitive function (unstandardized coefficient = .151, p < .001) had positive autoregressive effects over time. Participation had a cross-lagged effect on cognitive function (unstandardized coefficient = .061, p < .001). However, the cross-lagged effects of cognitive function on participation were not statistically significant (unstandardized coefficient = .051, p = .312). Various health-care programs that promote social participation and improve cognitive function must be established. Additional studies are required to confirm the causal effects of cognitive function on participation.