Abstract: In response to the challenges of wide fluctuation of density of medium suspension, unstable product quality, lagging of manual adjustments, and backward management level of traditional heavy medium separation systems, the Xiayukou mine coal preparation plant has developed an intelligent heavy medium separation system based on fuzzy control technology, in line with the plant′s specific conditions. The system adopts a three-tier architecture comprising equipment layer, control layer, and decision-making layer. It works with density deviation, density deviation change, and magnetic material content of medium suspension and qualified medium tank level as input data, and with openings of diversion valve and water supply valve as outputs, so as to realize intelligent control of heavy medium separation process through fuzzification of input data, fuzzy inference and precise computation of output results. Practice results demonstrate that application of the intelligent heavy medium separation system can lead to increase of clean coal yield by 1.43 percentage points and significantly enhanced density control precisions; the density fluctuation of the suspension has decreased from ±0.04 g/cm³ in the manual control mode to ±0.005 g/cm³; the control stability is higher than that in manual operations with the standard deviation of density decreased from 0.016 g/cm³ to 0.004 g/cm³; and because of no involvement of manual operations, the labor intensity is reduced. Application of the intelligent heavy medium separation system has not only realized unmanned and intelligent separation process, but also significantly stabilized product quality. The study made in the paper provides a valuable reference for similar coal preparation plants to realize intelligent transformation of existing heavy medium coal separation system.