Abstract: 【Purpose/significance】To validate the feasibility of applying a novel multi-agent collaborative governance model in online public opinion governance model and to provide empirical evidence【. Method/process】Initially, based on the multi-agent modeling ap⁃ proach, different agents and their attributes in online public opinion were defined. Subsequently, these agents were categorized into three models: network autonomy, traditional governance, and multi-agent collaborative governance, based on their roles in governing online public opinion. Finally, using real-world public opinion events, NetLogo simulations were conducted to analyze the evolution of online public opinion under each governance model【. Result/conclusion】The results indicate that in the network autonomy and tradi⁃ tional governance, it is difficult for network public opinion to achieve sustained positive outcomes; in the multi-agent collaborative gov⁃ ernance, netizens, media and government agents collaborate to form a synergistic effort, which can make the network public opinion get sustained and effective governance.【Innovation/limitation】The feasibility of the "multi-agent collaborative governance model" in gov⁃ erning online public opinion has been validated. However, the model assumes a static public opinion network, and its accuracy in rep⁃ resenting the real-world dynamics requires further refinement.