Abstract: Based on the statistical data of total water content,air-drying ash content,air-drying volatile and receired basis low calorific value of washed blend coal in Pingshuo Energy Sources Co Ltd coal preparation plant,using multiple linear regression method to establish mathematical model which is suitable for predicting the low calorific value of washed blend coal of the plant. This mathematical model can comprehensively check the report of quick ash and water tests. The establishing and perfecting of the prediction method has a certain guiding significance for coal quality monitoring.