Abstract: The life detection radar is developed for human life signals detection according to the doppler effect of body movements，breathing and heartbeat.It is widely used in rescuing survivors buried after disasters.Since respiratory signal has the characteristics of cyclical，and the echo signal of the undulating chest is strong，the doppler frequency shift is relatively easy to detect.According to the irregular respiratory frequency which lead to the changed thoracic cavity with multiple reflections，and the high noise background，the combination of multiple doppler frequency shift signals is established for the life echo model.In this paper，the method for detecting life information from low signal to noise ratio(SNR)echo signal by using MUSIC spectrum estimation is studied and the block diagram of the detection process is given as well.In the view of high SNR and low SNR，the FFT and MUSIC spectrum estimation methods are used to detect the life information on a radar range resolution unit.A number of tests in simulated conditions have shown the results as follows.In the high SNR echo，the FFT can detect the target，however，the life information of the low SNR echo spectrum is submerged by the noise spectral line；The MUSIC spectrum estimation method can suppress the noise spectrum，and the life information in the low SNR echo spectrum is also very obvious which can be detected.All of these results have indicated that the MUSIC spectrum estimation method can be used to detect the life information more reliably under the background of noise.