Abstract: To realize the real-time prediction of slag composition in the smelting process and provide the assistance to the operations in electric arc furnace (EAF) steelmaking process such as charging, the influence factors on the slag composition in the furnace (furnace reaction, charging, and slag overflowing) were studied, and the real-time prediction model of slag composition in EAF steelmaking process was established. In the results, the model could predict the slag quality, the slag composition, and the oxidation status of Fe element in the furnace in real time, providing the guidance for the auxiliary material charging and the slag flowing in the smelting process. Compared with the slag sampling results, the average relative errors of CaO, SiO₂, and FeO content in the slag between the actual measurement and the model predicted values were 12.66%, 11.17%, and 19.16%, respectively.