As a result of the imperfections of the automatic control of ore grinding in ball mills allowed electricity overrun balls and lining with a decrease in performance. The aim was to develop a computer adaptive control system capable to reduce the existing shortcomings. Theoreticallygroundedapproachreducingtheexistingshortcomings.It is suggested to go on grinding in ball mills in the mixture, and a certain type of ore, and control the process is not in fact there is a situation, and in particular identified an array of ore, which will be sent to a ball mill.. In a feed of mill with certain regularity accidentally changing the average size of the ore from which performance depends on both the boot and the optimum value of the liquefaction slurry. The developed computer adaptive control system has a present and a priori information support that allows you to create a defining impact on the estimation of the array cycle of ore at its load management, liquefaction of pulp, water consumption in the sand and on the surface of the original ore. In addition, the adaptive system with a sufficiently high accuracy determines a mass flow rate stabilizes at ore loading cycle, predicts and stabilizes the pulp dilution. Preliminary dilution sands single-spiral classifier wetting original ore and the remainder of spraying it with water for a given total amount of thinning the pulp in the central part to allow the inclusion of active initial section of the drum mill, which increases the grinding performance. Relay law formed by the same computing resources in the form of an ideal relay characteristic without hysteresis, and a low level of response, which may be 1% of the nominal value of the controlled quantity of the output device adaptive system are four similar automatic stabilization circuit parameters where used. Adaptive system reduces power consumption to 12%, the flow of balls and lining up to 7% and an increase in productivity of up to 8%.