In order to extract the characteristic parameters of gear crack fault and identify the types of gear with different crack depths,the vibration signals of gear with three different crack depths were collected from the cylindrical gear in the singlestage gear box.The collected signals were analyzed in timefrequency domain and decomposed in EEMD,the timedomain characteristic parameters and EEMD energy characteristic parameters were extracted,the fault characteristic vectors of gear cracks were analyzed and constructed,and the classification algorithm of support vector machine based on radial basis kernel function was used to identify gear with different crack depths.The results show that the fault eigenvectors of gear crack constructed by combining timedomain characteristic parameters and EEMD energy characteristic parameters can be used to accurately identify the types of gear with different crack depths.