Abstract: Photonic doppler velocimetry (PDV) is a new type of laser velocimetry system. It is widely used in the velocity measurement of the shock wave, detonation wave and other short-time high-velocity movement. Data processing is an important part of it, which purpose is to obtain the velocity information of the moving target from the measurement data mixed with a lot of noise. Based on the PDV principle, the sources and characteristics of noise about PDV data obtained from laser shock peening experiment are discussed. The de-noising principle of the fringe method is explained. Some unique methods are proposed to eliminate noise in wavelet transform. PDV data were processed using the fringe method, short time Fourier transform and wavelet transform. The errors, real-time processing and versatility about the three methods are compared. It is shown that wavelet transform is particularly suitable to PDV data processing of laser shock peening experiment.