Abstract: Aiming at the difficult problem of insufficient cutting control accuracy caused by complex geological structures and frequent changes in coal rock interfaces in coal mining working faces, an intelligent adaptive cutting system that integrates 3D geological modeling, multi-source perception, and fuzzy control optimization is constructed. Taking the actual measured data of No. 3 coal seam in Hecaogou No.2 Coal Mine as the basis, a linkage control mechanism for drum height adjustment, traction speed, and coal rock identification is established, and a load driven dynamic control logic is constructed; Multi-source information of visual, current, vibration, and pose are integrated to construct a coal rock perception model, a fuzzy rule base is designed to achieve automatic adjustment of drum height and collaborative control of traction speed, with a system response time of ≤ 100 ms. The on-site actual measurements show that this system decreases the power consumption by 17.5% per ton of coal, decreases the coal gangue content by 36.1%, and reduces the unit consumption of pick consumption by 34.6% (P<0.05), with excellent adaptive ability and economic performance, providing a reliable technical support for intelligent cutting control of medium-thick coal seams.