In order to enhance the level of time-frequency resolution with fault vibration signal,four kinds of time-frequency analysis algorithms,such as wavelet transform (WT),short time Fourier transform (STFT),synchrosqueezing transform (SST) and synchroextracting transform (SET),were used to analyze multi-component simulation signals and friction fault vibration signals of rotating machinery,and the principle of the SET algorithm was introduced in detail.The performance of SET was compared with other three algorithms through Renyi entropy.The results show that the SET algorithm can be used to effectively eliminate energy divergence,aliasing and endpoint effect for multi-component simulation signals,which is consistent with the real time spectrum of the original multi-component simulation signal,and the optimal time resolution and frequency resolution can be achieved;to friction fault vibration signals in field,the background noise in the friction fault signal can be suppressed well by using SET,the time-frequency diagram obtained has good time-frequency focus,and the energy divergence is greatly improved,so the corresponding frequency of the friction fault signal can be clearly seen in the time-frequency diagram.