Abstract: Semi-autogenous grind mill has many complex features such as multivariable，nonlinear，strong coupling，large lag，time-varying and so on，and many process parameters are difficult to be detected and control by conventional control method.Therefore，Wushan plant find out relationship among semi-autogenous grind mill feed ore granularity，mill noise，power，axial pressure，grinding concentration,feeding capacity during working，based on the human experience，and expressed by computer language，a qualitative，intelligent control system was obtained.Practice shows that semi-autogenous grind mill can control system according to the control strategy of the server settings，real-time acquisition the mill process parameters，automatically working on optimal parameters on semi-autogenous grind mill process，compared with the original manual control in Wushan Plant,the feeding capacity increased 24.7 t/h，prolong the service life of the lining board 11.1 d，mine energy consumption decreased 0.49 kWh/t，has significant economic benefits，has important application prospects in the field of metal mine.

Key words: Semi-autogenous grinding, Intelligent control, Mill noise, Feeding capacity