Aiming at the problems of motor bearing susceptibility to damage,long time consumption and low accuracy of traditional diagnostic methods,a motor bearing fault diagnosis method based on improved grey wolf optimization algorithm (IGWO) optimization support vector machine (SVM) was proposed.The ensemble empirical mode decomposition (EEMD) of motor vibration data was carried out to extract the IMF energy moment as the characteristic vector,combined with the IGWO-SVM classifier,the motor bearing fault detection model was constructed.Improved Tent chaotic mapping,nonlinear convergence factor and dynamic weight strategy were introduced into the model,and an improved classification algorithm was obtained,by which the optimal penalty parameter C and kernel parameter γ of the SVM could be found quickly and accurately.Through the experiment of motor bearing vibration data,the diagnostic results show that the accuracy of the bearing fault method is as high as 99.4%.Finally,the experiment verifies that the proposed diagnosis method has good algorithm stability and anti-noise performance,which can effectively improve the accuracy of fault diagnosis.