The pulverizing system serves an important function in the safe and efficient operation of thermal power units. The effect of coal moisture and pulverized coal moisture on the pulverizing system was considered in establishing a mass and energy balance-based dynamic mathematical model of a coal mill. The parameters of the model were identified by using the genetic algorithm and later validated with measurements. An extended Kalman filter was designed to estimate and verify the internal states of the coal mill. Based on the dynamic mathematical model of coal mill, pulverized coal moisture was used as a feed-forward signal, and a feed-forward compensation controller was designed to optimize the coal mill outlet temperature. Simulation results showed that the dynamic characteristic of the mill can be forecasted effectively with the established model. Moreover, pulverized coal moisture was reduced with feed-forward compensation, thereby improving the efficiency of the boiler. (C) 2014 Elsevier Ltd. All rights reserved.