This article presents the basis of a supervisory fuzzy expert controller for semi-autogenous grinding mill circuits. Stable feeding regimen to the mill, enhanced throughout, energy saving and human operator training are the most important objectives of this advanced control system. The fuzzy system calculates optimum set points for plant distributed control loops, causing them to tune semi-autogenous grinding mill performance to new operating set points. Although leading companies have their own commercial control packages, this supervisory controller is coded in MATLAB (R) and is able to connect to plant lower level controller. The controller has been tested and verified in Sungun copper concentrator semi-autogenous grinding circuit. Results proved the ability of proposed supervisory control system to increase the throughput of the mill by 3.26% and decrease the specific energy consumption by 6.29% at the same time. On the other hand, smooth set points calculated by the fuzzy control system, decrease the fluctuations in mill operation which finally results in a more stable operation of the grinding circuit. (C) 2018 Elsevier Ltd. All rights reserved.