Abstract: Ultra-Wide Band (UWB) radar exhibits strong penetration capability and high resolution, enabling the detection and localization of trapped personnel behind coal-rock collapses in mine disasters. This paper introduces the principles of UWB radar localization and its applications in mine rescue operations. The UWB radar personnel localization technologies are systematically reviewed from three perspectives: radar localization methods, static/dynamic target localization, and single/multi-target localization. Key challenges in mine rescue scenarios are identified: ① significant localization errors and limited effective detection range in thick, heterogeneous, and discontinuous media; ② weakened radar echoes and severe clutter interference under Non-Line-of-Sight (NLOS) conditions, leading to low-precision micro-motion target detection and large real-time errors for dynamic targets; ③ signal interference and occlusion effects among multiple targets degrading localization accuracy. Future research directions of UWB radar personnel localization technology for mine rescue operations are proposed: ① optimizing the UWB radar localization system by constructing cross-modal information fusion models and developing highly adaptive signal processing methods to enhance the system's adaptability to post-mining disaster environments; ② improving the applicability of combined static and dynamic target localization by developing hybrid localization algorithms that integrate Bayesian networks or deep belief networks to fuse static and dynamic target features and establishing state-switching-based comprehensive models; ③ improving UWB radar echo processing algorithms, combining adaptive beamforming technology, Multiple Input Multiple Output (MIMO) technology, and optimized K-means++ or entropy-based hierarchical analysis algorithms, effectively distinguishing multi-target position information, and validating their adaptability and reliability in complex environments through extensive simulation experiments.