The performance of existing rolling bearing fault feature extraction algorithms would be attenuated with the expansion of fault set.Aiming at the problems of interference and mode aliasing between fault signals,a feature extraction algorithm based on dual tree complex wavelet was proposed.Dual tree complex wavelet structure contained two independent filter banks,which formed a complementary relationship in the decomposition and reconstruction of noisy mixed signal,to improve the smoothness of signal sampling;the structure of dual tree complex wavelet filter bank was optimized to reduce the translation sensitivity of fault signal,threshold was used to process high-frequency wavelet coefficients to achieve the purpose of noise reduction,and the energy characteristics of the subband signal was extracted based on time series sample entropy.The experimental results show that the proposed algorithm can be used to accurately extract the fault feature information of each part of the rolling bearing,and the online fault recognition rate of the algorithm reaches 99.56%.