Abstract: This study focuses on the operational safety of urban rail transit and proposes a disaster early-warning and emergency response method that integrates multi-source sensing with intelligent analysis. Multiple types of sensors are deployed at critical monitoring nodes to achieve real-time monitoring of environmental and structural conditions, while multi-sensor data fusion enhances the completeness and accuracy of information acquisition. A fuzzy grey clustering algorithm is then introduced to process the collected data, and a risk evaluation index system is constructed according to disaster characteristics. By quantitatively calculating the influence factors of each index, comprehensive evaluation values are obtained and used to classify disaster risk levels. Based on the results of the background analysis, the system deeply integrates risk information with Building Information Modeling (BIM) technology to realize two-dimensional/three-dimensional visualization of disaster locations, surrounding emergency resources, and facilities. When the risk reaches a pre-warning threshold, the operation and maintenance platform automatically issues alarm notifications, enabling managers to quickly locate the disaster and dispatch personnel and materials. Meanwhile, the system links with firefighting facilities through the Internet of Things to perform automatic protective actions and intelligently plan optimal evacuation routes based on disaster scenarios. This approach enables rapid disaster identification, graded early warning, and coordinated emergency response, effectively improving the efficiency of rail transit emergency management and safety assurance, and providing key technical support for intelligent operation and maintenance.