Abstract: For washing the coal which comes from multiple seams and has varied and changing properties at Luotuoshan Coal Preparation Plant, the problems encountered are complicated coal blending calculations, high labor intensity and low washing efficiency. To tackle these problems, the particles swarm optimization(PSO)—based special mathematic software MATLAB is used to compile and develop the mathematic models. This makes it possible to predict the quantity-quality data and the coal separation parameters based on property of raw coal treated, make optimum calculation of the scheme for washing the different coking coal in blended manner, make rapid and accurate calculation of optimum blending ratios and provide the separation parameters for reference. Study result shows that through adoption of the scheme, while ensuring the quality indices of the clean coal products, there can be seen reductions in coal blending cost, labor intensity and cost, and much improved coal cleaning efficiency and economic result.