Safe evacuation is directly related to the safety of people's life and property, which is a very significant research topic. Based on the Bat Algorithm (BA) based crowd evacuation model, this paper comprehensively considers the influence of chaos degree on evacuation, introduces evacuation entropy for position correction, and constructs a BA evacuation model based on entropy correction. With the increase of the number of evacuees, the BA-based crowd evacuation model appears the phenomenon of queuing and congestion, which is alleviated by the model in this paper to some extent. Finally, the superiority of this model is verified by the comparison experiment with or without the entropy correction. The simulation results show that this model guides the crowd to avoid the chaotic area while advancing and effectively improve the evacuation efficiency.