Major chemical accidents may threaten the lives and health of people in the surrounding areas. Large-scale regional evacuation is a key measure for protecting the public from these accidents. In such unconventional emergency situations, the diffusion of evacuation warnings has a significant impact on the public's decision on whether to take evacuation actions. Through an investigation and analysis of evacuation cases, this paper discusses the diffusion characteristics of evacuation warning and evacuation efficiency of the public in major chemical accident cases. The results indicate that different diffusion methods of evacuation warnings affect public evacuation decisions. In addition, based on curve fitting and regression analysis, we propose a mathematical model of evacuation warning diffusion for responsive evacuation and an evaluation model of the relationship between the diffusion of evacuation warning and the evacuation rate. The results indicate that in the responsive evacuation process, the diffusion efficiency of the evacuation warning corresponds to Weibull distribution. As the diffusion of the evacuation warning continues, the evacuation rate of the people who receive evacuation warnings first decreases and then increases, when more than 50% of the people in the evacuation area receive the evacuation warning. The evacuation rate of the people who do not receive evacuation warnings increases first and then decreases. The results of the fitting analysis indicate that when -73% of the people in the evacuation area receive an evacuation warning, the area's overall evacuation rate is the highest. This can provide a basis and reference for regional evacuation analysis and emergency planning in unconventional emergency situations. (C) 2020 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.