Abstract: In order to further improve the intelligent control level of fully mechanized mining equipment in coal mines and achieve the normalized operation of intelligent mining face, this paper proposes the related concepts and system architecture of intelligent mining face driving control based on digital twin. The digital twin system of intelligent mining face is composed of physical equipment, virtual twin and virtual and real interaction model. The virtual model mapped with physical equipment is established by digital method to achieve the purpose of virtual and real interaction, intelligent decision, accurate control and dynamic evolution in fully mechanized mining equipment during operation. The paper expounds the normal operation requirements of intelligent mining face and related control difficulties of shearer, hydraulic support and scraper conveyor, and proposes the single-machine digital twin driving control architecture of smart mining face, including mechanism model, control model, twin data model and digital twin model synchronization and evolution. The method of digital twin virtual-real interaction is described, the Cyber-Physical System is used to guarantee the information interaction ability, and the knowledge model is used to solve the data congestion problem, which provides a real-time virtual-real interaction for the digital twin control system of intelligent mining face. A three-machine cooperative control method driven by the digital twin of intelligent mining face is proposed, including three-machine correlation, intelligent coal mining control, intelligent support control and intelligent transportation control. Finally, taking the practical application of digital twin in mines as an example, the intelligent control of digital twin is verified by establishing the relevant model and designing the relevant experiment scheme. By carrying out theoretical research on digital twin intelligent control architecture and mode, the paper aims to solve the current problems of low environmental perception, poor equipment prediction accuracy and high manual intervention intensity in fully mechanized mining face, realize an adaptive control and human-computer interaction of equipment under complex environmental conditions, and provide a reference for the intelligent construction of coal mines.