Abstract: To explore the effect of oxide composition on coal ash fusion characteristics, the oxide content in coal ash was selected as independent variables. On the platform of SPSS software, full subset regression and stepwise regression were carried out on deformation temperature DT, softening temperature ST, hemispheric temperature HT and flow temperature FT, respectively. The newly defined fusion index FI and the optimal regression prediction equation were obtained. The results show that the single oxide component has no significant effect on the ash melting temperature. The melt index with the most significant effect on DT is FID=Al2O3+Fe2O3, and when the FID content in coal ash is less than 30%, DT hardly changes, and when the content is greater than 30%, DT decreases greatly. The melt index with the most significant effect on FT is FIF=SiO2+Al2O3+Fe2O3, and as the FIF content increases, the flow temperature shows an upward trend. The melt index FIH=SiO2+Al2O3 which has the most significant influence on the hemispheric temperature HT, and the melt index FIS=SiO2+Al2O3+Fe2O3 which has the most significant influence on the softening temperature ST, but the significance of FIH and FIS on ST and HT is slightly lower. For a more accurate prediction model, the prediction equations of ST and HT are obtained by stepwise regression method analysis starting with ten oxides.