To develop a filtering algorithm which could trace the spectral peaks of the interference and which was used to extract the breath signal with the same band interference in radar non-contact life-detecting system, second filtering algorithm was studied. Through first filtering,the probable interfering spectral peaks (ISP) could be detected by Yule-Walker spectrum estimating and could be located by calculating the coefficients of normalized cross-correlation function according to standard breath signal. Thus the breath signal could be extracted through a second filtering. By using the second filtering algorithm (SFA), the same band interfering spectral peaks with breath signal could be recognized and inhibited. So we conclude that the same band mono-ISP could be inhibited by using SFA and breath signals could be effectively extracted.