Abstract: Ultra-wideband (UWB) radar echo signal processing affects the accurate detection of life signals, which is a key prerequisite for the success of mine rescue operations. Based on the echo signal processing process, this study introduces the working principle of UWB radar life information detection technology, discusses the relevant methods and algorithms of clutter suppression and background noise filtering in the echo signal preprocessing, and analyzes the research status of life signal extraction and recognition methods. It is pointed out that the present UWB radar life information detection technology has the following problems in the echo signal processing: filtering and denoising technology is not suitable for the environment; life signal extraction target is single and multi-method fusion technology needs to be improved; the generalization ability of human body recognition model is insufficient and the action recognition is not real-time. In view of the existing problems, the future development trend of UWB radar life information detection echo signal processing technology is proposed: the mine clutter model is constructed and the real-time adaptive filtering algorithm is studied. The extraction and fusion of multi-modal life signals and the optimization of data processing process are realized by using deep learning related technologies. The mining rescue body recognition technology based on multi-task learning and lightweight deep learning is studied.