A nonlinear robust control method for the air supply system of proton exchange membrane fuel cell was proposed. The task was to control the excess oxygen ratio and the stack cathode pressure to be not less than the expected value under the step load current to avoid oxygen shortage in the air supply system of fuel cell and hence obtain the maximum net power. A control-oriented air supply system model with multiple inputs and outputs was established. A nonlinear robust control method was designed based on the fact that the stack cathode pressure in the air supply system model could not be directly measured and the model had the characteristics of high-degree nonlinearity, strong coupling and much uncertainty of parameters. The method included a differential observer and a nonlinear robust controller based on feedback linearization. The simulation experiments were further carried out to verify the accuracy of the model. In addition, the robustness and effectiveness of differential observer and nonlinear robust control method for the fuel cell air supply system were verified under the conditions of step load current and uncertain system parameters.