Abstract: Abstract: In recent years, gas extraction has become an important means to control gas disasters. However, due to the limitation of sealing quality in the process of gas extraction, gas extraction is often affected by air leakage, which leads to the problem of low extraction flow and concentration. Therefore, in order to effectively solve this problem, this paper studies the intelligent control method of negative pressure in gas extraction borehole. Firstly, the theoretical derivation of ' gas-air ' mixed gas in coal seam is carried out. Secondly, based on the derived theoretical model, the gas drainage data under different initial parameters of coal seam are obtained by numerical simulation method and a large database of gas drainage is constructed. Then, four intelligent prediction algorithms are selected by using the large database and the BP algorithm with the highest prediction accuracy is determined. Then, according to the determination of BP algorithm, PSO algorithm is used to optimize and construct PSO-BP negative pressure prediction model. Based on this model, the intelligent control process of negative pressure of gas drainage borehole is put forward. Finally, industrial test is carried out in the field. The results show that BP algorithm is more accurate than ELM, TCN and SVM algorithm in extracting the characteristics of gas extraction data, and the prediction reliability of negative pressure of gas extraction borehole is higher. The optimization of BP algorithm by PSO algorithm can effectively improve the weak points of BP algorithm in dealing with complex data. The PSO-BP model constructed has a more significant improvement in the prediction accuracy of gas drainage borehole negative pressure than BP algorithm. The field test shows that after the intelligent control of the negative pressure of the borehole, the concentration of gas extraction is increased by more than 5 % compared with the concentration before the control, the pure amount of single hole extraction of gas extraction is increased by 0.015 m3 / min, and the pure amount of group hole extraction is increased by 0.023 m3 / min. The research results can effectively guide the on-site gas extraction work, improve the gas extraction effect, and ensure the safety of coal mine production.