Academic social network platforms are important means to carry out scientific research cooperation. This paper constructs a tripartite evolutionary game model of academic social network platforms, research team leader and researchers, describes the evolutionary path of the game and analyzes the sensitivity of parameters. The results show that initial strategy, benefits and costs of agents and platform rewards and punishments are important factors affecting system evolution. Finally, strategies to optimize the cooperative behavior of academic social network scientific research teams are proposed to promote scientific research cooperation and innovation.