The mass incident is an evolutionary game process involving multiple subjects. In this process, each subject has strategy space and payoff matrix in the state space. The subject can get different payoffs according to the feedback information from the environment and decide after comparing different payoffs. The entire process is called the evolutionary game of mass events. This paper will do the following research work: (1) analyzing the general process of mass incidents from the perspective of an evolutionary game; (2) constructing the evolutionary game model of mass incidents based on Q-learning; and (3) validating the model by using certain mass incidents.