Abstract: Flooding and drying failure seriously affect the working performance and service life of PEMFC. In order to fully characterize the high and low frequency electrochemical reaction information, wide-band PEMFC electrochemical impedance spectroscopy was established in this paper, and a water management fault diagnosis method based on second-order RQ-RLC equivalent circuit model was proposed. First, a fuel cell test bench was built and water management fault simulation experiment was conducted to obtain the corresponding electrochemical impedance spectrum, identify the parameters of the second-order RQ-RLC equivalent circuit model, and obtain an eight-dimensional water management fault data set. Then, the linear discriminant analysis method was used to reduce the dimension of the high dimensional water management fault data set to obtain the fault feature sample set, and four key parameters of the model were selected as the fault diagnosis feature parameters. Finally, the adaptive differential evolution optimization support vector machine algorithm was proposed to classify fault feature sample sets, the test set classification accuracies of 100%, 97.44% and 95.24% at 50, 130 and 210 sets of samples, respectively, showing that the proposed method can accurately diagnose the type of water management fault the fuel cell is subjected to.