Abstract: Aiming at problem that fully mechanized caving face at the back end of fully mechanized top coal caving face adopted manual caving control method, its automation level differed greatly from that at the front end of fully mechanized top coal caving face, an automatic coal caving control system was studied. According to inclination sensors installed on top coal caving hydraulic support and travel sensors in jacks of front or rear scraper conveyor, position and attitude of caving mechanism of the hydraulic support are controlled to achieve precise control of caving quantity; by installing pan-tilt camera and infrared transmitter on tail beam of the hydraulic support, real-time video monitoring is carried out in whole process of coal caving to realize detection and control of coal transportation quantity and straightness of the rear scraper conveyor; image gray recognition technology is used to identify coal-gangue and detect gangue content of top coal. The system improves efficiency and quality of coal caving, and provides an idea for realizing unmanned fully mechanized top coal caving face.