Abstract: In response to the theoretical and technical challenges faced by mine intelligent safety, this study constructs a safety intelligent system and an overall solution of mine intelligent safety.Firstly, it analyzes the core issues of safety intelligence, pointing out that the fundamental contradiction lies in the lack of safety theories adapted to intelligent environments.Secondly, a five-layer safety intelligent system is established, covering theories, methods, technologies, equipments, and systems.It emphasizes the integration of system fault evolution theory (to grasp overall trends) and large language models (to analyze detailed correlations) to construct the intelligent safety theories, with data flows running through field equipment, information platforms, and core systems.Finally, targeting the three main threads of mine natural disasters, equipment failures, and human-caused accidents, an overall solution is designed that includes a theoretical layer (system fault evolution theory and large models) and a technical layer (intelligent equipment and information systems), achieving a closed-loop management scheme for data collection, situational awareness, and decision-making coordination.This study provides a complete framework consists of theories and applications for intelligent mine safety, highlighting that the deep coupling of intelligent theories and safety principles is the key to realiz mine safety.