For the current status of research on the factors influencing CSR communication in users' participation in social media, which is dominated by information content features, this study combines corporate features with information features based on ELM. In this study, 4183 CSR communication data obtained from Sina Weibo were studied using negative binomial regression analysis. The results prove that sentiment of messages and message content quality such as information presentation richness, interactivity, CSR tags, CSR quantification results and CSR COVID-19 pandemic theme influence user participation from the central route, while enterprise features such as controversial companies and number of followers influence user participation through the peripheral route. Specifically, information presentation richness, interactivity, CSR quantification results, number of followers and sentiment of messages positively influence user participation. CSR tags and COVID-19 pandemic theme positively influence users' reposts. CSR tags positively affect users' likes. Controversial corporate background negatively influence users' reposts and comments, and positively influence users' likes.