In the process of mine construction and production, mine water hazards often cause heavy casualties and property losses, and even environmental pollution problems that are difficult to recover. To prevent such accidents, various countries and mine enterprises have paid great attention to them and based on conventional prevention and control methods, they have continuously tried to introduce new technologies and new methods such as informatization and intelligence. This makes the hydrological data generated during mine mining explode, but these data have never been fully and effectively used. Therefore, how to use mine hydrological data to better serve the safety of mine water inrush has become an important breakthrough in mine water inrush prevention. A mine water inrush safety guarantee system based on machine learning is implemented to realize a mutually connected system closed-loop, timely warn mine water inrush accidents, and recommend targeted solutions. The technical support for multi-angle is comprehensively provided to systematically prevent and control mine water inrush accidents.