In this study, the operating condition (operating temperature, anode pressure, cathode pressure and current density) and channel structure (heights of channel inlet and outlet) of a proton exchange membrane fuel cell (PEMFC) are optimized using multi-objective genetic algorithm. The optimizations of the operating condition and channel structure are based on a PEMFC stack model and a three-dimensional, steady-state, non-isothermal PEMFC model, respectively. The optimal operating condition and channel structure are selected by TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) from the optimal solution set called Pareto front. After the optimized operating condition is obtained, it is applied to the optimization of the channel structure. The results present that the optimal channel structure under the optimal operating condition is a type of tapered channel. Compared to the conventional straight channel, the tapered channel can enhance gas reactant transport in the PEMFC and a greater amount of reactant can participate in the electrochemical reaction, thus, more output power can be obtained. (C) 2017 Elsevier Ltd. All rights reserved.