Abstract: In order to improve the safety and efficiency of gas extraction and reduce the economic cost of gas extraction,the safety constraints and efficiency constraints of gas extraction system operation are analyzed. The four control tasks of gas drainage system are analyzed and the mathematical model of gas drainage optimization. According to the theoretical control strategy,the complete process of intelligent control of gas extraction is put forward. On the basis of the above regulation process,an intelligent regulation model of gas extraction is proposed,which takes gas extraction concentration,gas extraction pure quantity,gas extraction negative pressure and extraction pump efficiency ratio as the controlled quantities,and the valve opening of extraction drilling hole and extraction pump power as the controlled quantities. The simple RNN is used to analyze and process the time-varying law of the historical data of the controlled quantities,and learn the ideal dynamic fitting curve of the controlled quantities changing with time. The model predictive control algorithm (MPC) is used to intelligently control the controlled variable,so that the actual value of the controlled variable infinitely approaches the reference value at the corresponding time of its ideal dynamic fitting curve. Using correction feedback and rolling optimization,the anti-interference ability of intelligent control model of gas extraction is continuously enhanced,and finally the safety and efficiency of coal mine gas extraction are improved. Taking the simulated gas extraction data as an example,the algorithm simulation experiment is completed. The experimental results show that the overall change trend of gas extraction concentration decreases with time from 40%-5%,and the overall change trend of pure gas extraction quantity decreases with time from 9.0-5.0 m3/min. The ideal dynamic fitting curve obtained by cyclic neural network has a good data fitting degree,which can accurately reflect the change law of gas extraction concentration data and pure gas extraction data. What’s more,the negative pressure of gas extraction and the efficiency of gas extraction pump can be accurately maintained between 10-30 kPa and 1.3-1.5 m3/(kW·h),which meets the economic and safety needs of gas extraction process. The model predictive control algorithm can overcome the interference of environment and nonlinear factors to achieve better control effect，which provides a certain reference for intelligent control of gas extraction.