With the rapid development of radar technology, the radar target signal extraction under low signal to noise ratio (SNR) is becoming more and more important. The traditional pulse compression algorithm has no obvious effect under low SNR. A new algorithm for radar target echo signal extraction under low SNR is proposed. This algorithm is based on the traditional pulse compression algorithm, and use the wavelet transform to decompose the signal after pulse compression, then we extract the wavelet coefficients of the high frequency information. After that the high frequency information of the signal is denoised by the improved wavelet transform modulus maxima algorithm, and then the high frequency information of the signal is reconstructed. Finally, the complete target echo signal is reconstructed. The simulation results show that compared with the pulse compression and other algorithm, the performance has been greatly improved.