Abstract: In the context of the rapid development of information technology， social networking platforms are constantly emerging and widely used. While it brings convenience to life， it also leads to the emergence of public opinions on social networks. How to effectively deal with public opinions on social networks is a new challenge facing the government. After the multi-agent factors with high degree of relevance to public opinions on social networks are analyzed， Vensim PLE software is used to construct the multi-agent coping model for public opinions on social networks， and the specific case of “Yabuli Incident” is used for simulation analysis. The results show that public opinions on social networks are affected by the public opinion events， the media， the netizens， and many government entities and related factors. The government can cope effectively with public opinions on social networks through such measures as improving the ability to control public opinions， regulating and guiding media behaviors， actively channeling netizens' emotions， and improving information transparency， etc.

Key words: system dynamics, public opinions on social networks, social networking platform, simulation analysis