Abstract: It is difficult to meet requirement of field production by theoretically calculating fixed ratio of raw coal in blended coal preparation process of coal preparation plant, and problems such as large randomness, low quality of clean coal and high labor intensity are existed in raw coal ratio adjustment according to manual experience. According to above problems, an intelligent ratio control system for raw coal in coal preparation plant was designed. Least squares support vector machine is used to establish a prediction model of intelligent raw coal ratio, and particle swarm optimization algorithm is adopted to optimize model parameters. The system takes measured ash content value of raw coal, measured sulfur content value of raw coal, average feed amount of raw coal per hour, separation density, measured ash content value of clean coal, ash content target value of clean coal, measured sulfur content value of clean coal and sulfur content target value of clean coal as input variables of the model, so as to obtain predicted value of raw coal ratio. Coal feed amount of coal feeder is measured by belt scale to calculate measured value of raw coal ratio, and deviation of raw coal ratio is obtained by comparing the measured value with the predicted value. PID controller controls frequency converter of the coal feeder according to the deviation amount to achieve accurate adjustment of raw coal ratio. The actual application results show that fluctuation range of ash and sulfur content of clean coal is significantly reduced and quality stability of clean coal is good. Difference between ash content of clean coal and ash content target value is controlled within ±0.2%, and difference between sulfur content of clean coal and sulfur content target value is controlled within ±0.15%, which verify quality improvement of clean coal.