An appropriate operating temperature is crucial for maintaining the high efficiency and stability of proton exchange membrane fuel cells (PEMFCs). However, due to the strong nonlinearity of PEMFC systems, the application of classical PID control faces numerous challenges. Therefore, a radial basis function (RBF) neural network is constructed in this paper to achieve system identification and provide Jacobian information of the system, which is then integrated with a self-correcting PID algorithm to develop an RBF-PID temperature controller. Subsequently, based on the controller’s operational principles, six key parameters (including inertia weight and learning rate etc.) that determine its performance are selected for single-objective optimization using particle swarm optimization (PSO) algorithm. After optimization, both the overshoot and settling time of the control system are significantly improved. Finally, building upon the parameters of the PSO-RBF-PID controller, a multi-objective optimization is performed using the non-dominated sorting genetic algorithm III (NSGAIII), with the integral of time multiplied by absolute error (ITAE), average overshoot (), and integral of absolute identification error (IAIE) as the three objectives. The resulting NSGAIII-RBF-PID controller further enhances the control performance and achieves optimal temperature control, effectively expanding the applicability of PID controllers in the field of PEMFC temperature regulation.

NSGAIII multi-objective optimization, PEMFC, PSO, RBF-PID, Temperature control system