The invention discloses a coal ash fusion point prediction method based on a BP network model. The coal ash fusion point prediction method includes the steps: based on blast furnace injection pulverized coal ash, a three-layer BP network model with oxide components of coal ash as input and coal ash deformation temperature as output is constructed; and after a sample is determined, all parameters of the BP network structure are gradually determined through online training of the sample, and finally an optimal solution is obtained through application of a BP network model. The coal ash fusion point prediction method has the advantages that 1) the BP network model adopts an error back propagation mode, and the purpose of minimizing the error value is finally achieved through step-by-step adjustment of errors; and 2) due to the fact that measurement of the fusion point of the blast furnace injection pulverized coal ash has certain timeliness at present, the fusion point of the coal ash canbe rapidly and accurately predicted through establishment of the network model.