Abstract: In order to overcome the limitations of traditional prediction methods，a new model that can accurately predict the pollution characteristics of high-alkali coal combustion was provided and a proper range of temperature was given according to the pollution characteristics of coal.This prediction model adopts a three-step chemical extraction method to to detect the sodium content in the coal sample to avoid the sodium volatilization loss caused by coal ashing.Therefore， the total sodium content and various forms of sodium in the coal sample can be directly and accurately determined.In addition， this prediction model not only considers the effects of sodium， potassium， and solid-phase pollution sources， but also considers the effects of other volatile components (such as alkaline earth metals， sulfur， and chlorine) on coal pollution characteristics， and introduces temperature into the pollution.The prediction and quantification interval of the staining index RT obtained according to this model is as follows:when RT≤0.60，the contamination characteristics of coal is weak；when 0.60＜RT≤1.00， the contamination characteristics of coal is medium； when RT＞1.00， the contamination characteristics of coal is strong.Shaerhu coal and Shenhua Zhundong coal are used to verify the accuracy of this prediction model， the results show that:①At 700 ℃， RT=0.59＜0.60， Shaerhu coal is in a weak contamination area； at 850 ℃，0.60 ＜ RT= 0.83 ＜ 1.0， Shaerhu coal is in a medium contamination area； at 900 ℃，0.60 ＜RT=0.95＜1.0， Shaerhu coal is in a medium contamination area； at 950 ℃，RT=1.11＞1.0， Shaerhu coal is in a strong contamination area.②Shenhua Zhundong Coal has weaker contamination properties than Shaar Lake coal.At 850 ℃，RT=0.53＜0.60， ShenhuaZhundong coal is in a weak contamination area； at 900 ℃，0.60 ＜RT=0.78＜1.0， Shenhua Zhundong coal is in a moderately contamination area； at 950 ℃，0.60＜RT=0.87＜1.0， Shenhua Zhundong coal is in a moderately contamination area.The prediction results are consistent with the combustion test results of Shaerhu coal and Shenhua Zhundong coal， respectively， further verifying the accuracy and reliability of this prediction model.This prediction model can be applied to the prediction of contamination characteristics of high alkali coal.