Abstract: The tunnel excavation is a key and pioneering project in coal mining, and its human-machine-environment system exhibits characteristics of multi-disaster coexistence and coupling. It forms a special risk system featuring the coexistence of multiple disasters such as water inrush, gas, roof collapse, and mechanical injury, making the safety risk management extremely challenging. In response to the demand for intelligent construction in coal mines, a multi-disaster risk discrimination model based on the coupling of human-machine-environment systems has been established. A technical path for proactively suppressing disasters based on information perception and intelligent decision-making has been proposed. An intelligent safety risk control and emergency platform for tunneling faces has also been developed. The collaborative monitoring of risk information for the tunneling workface has been achieved, addressing issues such as poor coordination among the tunneling equipment group under complex conditions, lack of information connectivity, and insufficient intelligent control measures. This platform is based on multi-source heterogeneous data fusion (Service-Oriented Architecture architecture) and digital twin technology, integrates core technologies such as disaster risk prediction and warning (based on Transformer time series model) and intelligent video analysis of personnel violations and dangerous states (based on the “YOLOv5 + Long Short-Term Memory Network” model), and achieves: Three-dimensional visualization identification and dynamic warning of main risks such as water inrush, gas, roof, and mechanical injury; Deep correlation analysis and closed-loop management of multi-source data of human-machine-environment-management; Intelligent linkage regulation of ventilation and dust removal equipment; AI intelligent recognition of personnel violation behaviors (accuracy rate > 85%). The application of this platform in Ningtiaota Coal Mine of Shaanxi Coal Industry Group has demonstrated that it significantly improves the safety risk management capability of the tunneling face, and achieves a risk warning accuracy rate of over 92%, effectively responding to the requirements of the “Opinions on Further Strengthening Mine Safety Work” for disaster chain prevention and control, and providing key technical support for the intrinsic safety of tunneling in intelligent coal mine construction.