Abstract: The intelligent fully-mechanized mining face is an important component of the intelligent development of coal mines. However,the amount of intelligent mining practice data based on geological models is relatively small,and the model accuracy is limited,which cannot effectively utilize the historical data of coal shearer cutting,and the cutting efficiency is relatively low. Therefore,this paper proposes a method for planning the cutting path of coal shearer based on machine learning algorithms. Applying the historical exposed data in the process of cutting path planning can effectively improve the utilization rate of historical cutting data,and achieve precise cutting of the coal shearer. We obtain the following understandings and achievements: The simulated cutting path was verified using the height data of the drum before the cutting of real coal shearer. The results proved that the predicted coal seam thickness fluctuation pattern of the mining stage model was relatively consistent with the real coal seam thickness fluctuation pattern,which could ensure the intelligent mining accuracy of two shifts. The combination of 3D geological models and artificial intelligence has further enhanced the intelligent level of coal mining machine cutting,laying a solid foundation for truly achieving unmanned operation in planned cutting.