In order to solve the problem that gas emission from coal mining face can not be predicted accurately in real time, the difficulties and treatment methods of gas content inversion in working face are expounded by analyzing the sources and influencing factors of gas emission from mining face. On this basis, the inversion model of gas content is established, and the dynamic prediction method of gas emission from working face based on gas content inversion is put forward. The prediction method is based on the premise that the occurrence state of gas in coal is a continuous change, and the amount of gas emission is predicted through the inversion of coal gas content in working face and the determination of gas emission characteristic coefficient. The industrial test in a coal mine in Shaanxi shows that the prediction method can better predict the amount of gas emission from the coal face. The absolute error range between the predicted value and the measured value of gas emission is between 0.09~0.20m3/min, and the error percentage is controlled within 20%. The application of this prediction method can make real-time and continuous prediction of gas emission from mining face, ensure the safe and efficient production of coal mine, and provide technical support for the prevention of coal and gas outburst and gas overrun.