Abstract: To solve the problem that the existing methods have a poor effect on narrowband interference suppression in partial discharge(PD) signal, the method based on generalized S transform and fast independent component analysis is proposed in this paper. Firstly, the generalized S-transform is used to analyze the time-frequency distribution of the noised PD signal. In the time-frequency distribution picture of the PD signal, the characteristic region of the narrowband interference can be determined according to the different time-frequency characteristics of the narrowband interference and the PD signal. Then, in the characteristic region of narrowband interference, the frequency of narrowband interference is estimated by using the Candan algorithm. Finally, the PD signal is separated by using the frequency estimation of narrowband interference and the fast independent component analysis method, suppressing the narrowband interference. The simulation and measurement results show that the proposed method can effectively suppress the narrowband interference in the noised PD signal, and it can accurately obtain the waveform characteristic of the PD signal. Compared with the singular value decomposition method and fast Fourier transform filtering method, the proposed method has a better suppression effect on narrowband interference and less residual interference energy.