Because of the noteworthy characteristics of ultra-wideband (UWB) radar, it has been applied in the noncontact measurement of vital signs. The fluctuating movement of the chest caused by respiration and heartbeat of the human body is comprehensively reflected in the radar echo. A key problem is how to separate respiration and heartrate signals in the echo then evaluate the vital sign parameters. In this study, the position of a monitored object is judged from the output data frame of the radar, then the data sequence of chest movement in the time domain is extracted. The signal is filtered by the wavelet threshold method using the sym6 wavelet basis function, then the spectrum of the denoised signal is analyzed. This method can effectively extract the respiration rate and heartrate and achieve real-time monitoring. Its feasibility and effectiveness were verified by several sets of experiments, which indicate that the method can be applied in practice. This method is expected to help popularize the use of UWB radar in medical monitoring.