Abstract: To solve the problem of low accuracy of fault diagnosis of the high-pressure fuel pump based on time domain features,a feature extraction method for parameter optimization based on variational mode decomposition（VMD） algorithm and dispersion entropy was proposed,and support vector machine（SVM） was used for fault diagnosis. Firstly,based on the analysis of the working principle and typical faults of the high-pressure fuel pump,a simulation model of the high-pressure fuel pump built on AMESim platform was used to simulate the faults and collect the signals. After that,to solve the issue that the decomposition effect of VMD is limited by the number of decomposition and the selection of penalty factors,improved grey wolf optimization（IGWO） algorithm was used to optimize the parameters. Then,the dispersion entropy value of each intrinsic mode function was calculated to form the fault feature vectors. Finally,SVM was used to train and diagnose the fault feature vectors,and the fault diagnosis of the high-pressure fuel pump was realized. The results show that the fault diagnosis accuracy can reach more than 95%,and the method proposed is effective to the fault diagnosis of high-pressure pump.