Abstract: With the continuous advancement of intelligent mining, fully mechanized mining working faces still face challenges in equipment collaboration, data interoperability, and real-time remote monitoring. To address these issues, this paper proposes a digital twin simulation method for fully mechanized mining working faces based on Unreal Engine （UE） and designs a layered system architecture. The system consists of five layers: the physical layer, perception layer, communication layer, platform layer, and presentation layer. Multi-source sensors are deployed to acquire equipment operation and environmental parameters in real time. The collected data are transmitted through 5G and WebSocket channels to the platform layer, where MySQL and MemoryCache ensure efficient data cleaning, processing, and storage. At the presentation layer, a Web interface and a UE5.3-based 3D simulation module are used for visualization and interaction. Key innovations include high-fidelity 3D modeling, dynamic physical simulation with the Chaos engine, blueprint-driven interaction logic, and Pixel Streaming technology, enabling low-latency synchronization between physical operations and virtual scenes. Application results show that the system achieves dynamic monitoring and intelligent warning of key equipment, significantly improving operational efficiency, situational awareness, and remote maintenance capabilities. This research provides a new approach for constructing high-fidelity, interactive, and real-time digital twin platforms for coal mining faces, laying a solid foundation for intelligent and unmanned mining.