Abstract: Gas drainage is the most effective method to prevent gas accidents. In the process of drainage, the negative pressure of drainage has a crucial influence on the effect of gas drainage. In order to improve the effect of gas drainage, many mines use artificial regulation to adjust the negative pressure of drainage. However, artificial regulation often depends on artificial experience, and there is a problem of inaccurate regulation. With the development of intelligent technology, intelligent regulation technology is considered to be an important method to solve the above problems. In order to effectively solve the problem of inaccurate negative pressure regulation in the process of gas drainage, this paper uses three methods: numerical simulation, intelligent algorithm construction and field test to study the dynamic regulation strategy of intelligent gas drainage negative pressure. In the aspect of numerical simulation, the permeability evolution model in front of coal mining face is constructed. Three negative pressures of 13 kPa, 18 kPa and 25 kPa are set up, and the relationship between the negative pressure of drainage and the concentration and flow of drainage is studied and analyzed. In the aspect of intelligent algorithm construction, the LSTM prediction algorithm is constructed to predict the negative pressure of drainage. Based on this, the dynamic regulation strategy of intelligent drainage negative pressure is proposed. In the field test, a field test with a 120 days is carried out. The dynamic regulation strategy of intelligent drainage negative pressure is verified by installing intelligent regulation valves and monitoring equipment in different areas. The results show that the permeability of the pressure relief zone in front of the coal mining face is the largest, the permeability of the stress concentration zone is the smallest, and the permeability of the original stress zone is not affected by the mining disturbance. There is a positive correlation between the negative pressure and the gas drainage flow, and a negative correlation between the negative pressure and the gas drainage concentration. In the stress concentration area, when the drainage negative pressure is 25 kPa, the permeability no longer plays a leading role in gas flow. At this time, the pressure difference between the gas pressure in the coal fracture and drainage negative pressure plays a leading role in the gas flow. The field application effect of the dynamic regulation strategy of intelligent gas drainage negative pressure based on LSTM prediction algorithm is better. The gas concentration in different zone has been significantly improved after intelligent regulation. The research results of this paper have played a positive role in promoting the development of gas intelligent drainage technology.