Abstract: It is a more convenient way to use vibration signals for the fault diagnosis of diesel engine since such signals contain a lot of useful information which can reflect the status of the diesel engine. In view of the non-stationary characteristics of the vibration signals, the empirical mode decomposition was used to decompose the signals obtained, the three main IMFs of signals were selected approximately to replace the original signals, and their energy percentage, gravity frequency, center of gravity amplitude, skewness and kurtosis were used as the feature vector of the status of the diesel engine. Based on support vector machines, the diesel engine fault diagnosis was conducted applying vectors obtained in the method presented in this paper. Diagnostic accuracy rate reached above 92%, which verified the feasibility of the method. The research can provide a reference for other mechanical equipment fault diagnosis.