Abstract: A thermal management control method (APSO-BP-PID) with adaptive particle swarm optimization algorithm (APSO) to improve the control method of back propagation neural network proportional integral differential (BP-PID) is proposed to address the issues of large temperature fluctuations and poor response speed in the high-power proton exchange membrane fuel cell (PEMFC) thermal management system of city buses under continuous load changes and changes in operating parameters. This improves the slow learning speed and susceptibility to local extremum problems of BP-PID, enable the fuel cell system to quickly adjust and reduce temperature fluctuations when operating conditions change. Building a model simulation on the Simulink platform, using the proposed method to control the outlet temperature and inlet/outlet temperature difference of the stack, and comparing it with two control methods, BP-PID and PID, the results show that the APSO-BP-PID method has better control effect. Compared with BP-PID and PID, the average adjustment time under continuous variable load conditions is shortened by about 59 seconds and 97 seconds, respectively, and the temperature fluctuation is relatively reduced by 46% and 40% when the working parameters change. The proposed control method has smaller temperature fluctuations and shorter adjustment time.