Since hematite ore exhibits low grade, fine-grained and non-homogenous distribution, fluctuations occur frequently in its key variables during regrinding process. These key variables include the flowrate and the density of slurry produced in the first stage grinding, the underflow of the hydrocyclone and the dilution water flow added to the sump. Such fluctuations would generally lead to large variations in the sump level and the pump speed would inevitably vary in a wide range when using existing setpoint control for the sump level and the grinding density. Therefore, the classification efficiency of hydrocyclone would be reduced significantly. In this paper, a hybrid intelligent control (HIC) method is proposed, which includes a fuzzy switching controller of the sump level interval, a multi-PI switching controller of the hydrocyclone feeding pressure and a conventional controller of the hydrocyclone feeding density. By switching between a retainer and a fuzzy compensator to hydrocyclone feeding pressure setpoint, it has been shown that the sump level interval controller can guarantee the variations of feeding pressure setpoint to be within its desired range, and the multi-PI switching controller of the feeding pressure can track its time-varying setpoint. Thus, the sump level and the oscillations of the feeding pressure can be both controlled within their target ranges, respectively, and the feeding density can track its setpoint. The proposed method has been successfully applied to a regrinding process in a large-scale hematite beneficiation plant. As a result, the safe operation of the regrinding process has been realized and the qualified rate of the particle size index has been effectively improved.