Aiming at the phenomenon of high probability and lack of effective identification methods of mechanical failure of induction motors,a support vector machine classification model was proposed based on time-frequency entropy features.By building a fault simulation platform,the vibration signals were collected for multiple types of samples such as normal operation,dynamic eccentricity,misalignment,loose base and bearing faults,multi-dimensional statistical indicators were extracted,and the feature selection method was used to reduce the time complexity and ensure the accuracy and timeliness of the diagnosis.The model training was performed by combining with the support vector machine to complete the fault diagnosis.The experimental results show that the proposed method has high accuracy and good consistency in the existing sample data,and the overall method is simple to implement.