Abstract: Aiming at a series of technical problems,such as stronger mine pressure in high-strength mining face, and difficut prediction accurately, as well as poor surrounding rock control method, an intelligent working face rock disaster early warning strategy based on dynamic state perception of roof state, real-time identification of slice formation and multi-source monitoring information of stope is proposed. An intelligent warning mechanism for surrounding rock disasters in working face, real-time monitoring and dynamic sensing technology for top surface of working face was developed, and the pattern recognition of the position, duration and pressure strength of roof was also realized;mine pressure data based on big data platform was developed. Intelligent self-learning technology, combined with rock stratigraphic map can invert faults, structural formation and instability of key layers of the overlying strata in real time, and achieved accurate prediction of large-area pressure on top surface of working face; the research and development of coal wall stability based on 3D laser scanning can achieve high-precision monitoring of displacement of coal wall and the evolution process of joint fissure. It can also quantitatively describe shape, position and strength of rib after instability of coal wall, overcoming the problem that large areas of the previous working face were pressurized and coal wall spalling cannot be accurately predicted. This further integrated the factors such as coal cutting machine cutting speed, the working resistance of support and the multi-source monitoring information of the stope volume and other fields, and an intelligent sensing system was constructed, which combines the working face mining process with coal wall rib,unbalanced gas emission, and large-area instability and degeneration of the roof. The technical idea has been successfully verified in No.15106 intelligent super-long working face of the Sijiazhuang mine of Yangmei Group, which guarantees the safety of working face.