Abstract: Using the waveform characteristics of the valid values by dq conversion, wavelet transform and FFT, seven types of power quality disturbances are layered recognized, and their characteristic parameters are obtained.According to the distribution characters of the valid values of the disturbance signal, the disturbances are divided into two types. The first type includes voltage sag, swell and interruption. The second type includes harmonic, transient pulse, low frequency oscillation and voltage fluctuation.The multi resolution analysis is carried out for the second disturbances. The transient pulse and low frequency oscillation are identified by the characteristics of the high frequency coefficients, and then they can be separated by the number of zero crossing points.After FFT transform, harmonic and voltage fluctuation, whose high frequency coefficients distributions are similar. They can be identified according to spectrum characteristics. The proposed method can effectively identify the disturbance types and obtain their characteristic parameters, such as the starting and ending time and amplitude of voltage sag, the peak and peak times of transient pulse and low frequency oscillation. Simulation results show the effectiveness of the proposed method.