Abstract: To address the issues of isolated elements, static and rigid models, and insufficient integration of prevention and control engineering in traditional gas geology research, a "one-core, three-dimension, four-level" framework for constructing a gas geology system of a transparent working face is proposed. Centered on an integrated dynamic model and incorporating the three dimensions of geology, gas, and engineering, the system establishes a progressive technical architecture comprising the data layer, model layer, platform layer, and application layer. In the data layer, using a hybrid relational and non-relational database architecture as the foundation, standardized integration of multi-source heterogeneous data is achieved through spatiotemporal alignment, collaborative correction, and multi-dimensional fusion. In the model layer, based on the fused data and driven by three synergistic mechanisms—mining progress, detected anomalies, and prediction deviations—a digital twin model is constructed to enable dynamic coupling and adaptive updating of gas, geology, and engineering data. In the platform layer, leveraging the dynamically updated digital twin model, a 3D visual interactive platform that supports multi-dimensional spatial quantitative analysis is developed. In the application layer, the system integrates key application scenarios, including precise prediction of gas occurrence, 3D inversion and intelligent borehole design, anomaly identification, gas emission early warning, and closed-loop evaluation of prevention and control. The system achieves continuous parameterization and spatial quantitative expression of gas geological conditions in the working face, providing reliable data and decision support for precise disaster prevention and control as well as dynamic engineering deployment in high-gas and outburst coal mines.