Abstract: Based on the existing coal of a thermal power plant in China, taking the calorific value, volatile matter, ash, moisture and other indexes of coal as constraints and the coal blending cost as objective function, the coal blending optimization model of power generation enterprises is established. According to the linear programming method and the random direction method, the MATLAB software should be used to solve the problem, and the economic coal blending proportion of the existing coal of the enterprises meeting the constraints is obtained, It provides a new idea for power coal blending. The results show that the calculation results of the two methods are similar, and the economic coal blending proportion and the lowest cost can be obtained. The results of random direction method are slightly better than that of linear programming method. At the same time, the results can provide a reference for the purchase of coal in power plant and guide the inventory of power plant in advance.