Abstract: The paper presents a novel method which combines HHT-Hilbert Huang Transform with wavelet denoising to solve the difficulties for accurate locating of transient power quality disturbance in condition of severe noise pollution. The process of the method is that the signal is processed by applying Hilbert phase shifting to obtain a noised envelope of disturbing signal, and then the noise signal is removed by using wavelet denoising to get a disturbance signal with accurately located at start-stop time of the disturbance, and the accurate location of start-stop time of the disturbance can be obtained by disturbance localization algorithm. The simulation results show that this method can replace the lifting wavelet method to give an accurate location under severe noise condition. Compared with the method of extraction after denoising, it was concluded that the method of denoising after extraction has higher precision and better stability.