<p>Sintered ore is a vital raw material in the ironmaking process conducted in blast furnaces. The regulation of its alkalinity significantly affects slag emissions, energy consumption during iron production, and the overall efficiency of the blast furnace. This study proposes a comprehensive and reliable approach for managing the alkalinity of sintered ore through the development of a predictive model that employs a multi-channel neural network (MCNN) in conjunction with the informer framework, which incorporates multi-scale feature fusion. Initially, the maximum correlation coefficient (MIC) method is employed to identify variables that exhibit a strong correlation with the alkalinity of sintered ore. Subsequently, the MCNN is utilized to extract multi-scale features from the input variables, resulting in the generation of multiple feature maps that are subsequently concatenated and fused. The processed data is then fed into the Informer model to forecast the alkalinity of sintered ore. Verification has been conducted using empirical data from a sintering plant, and the performance of the proposed model is compared against four other models: LSTM, XGBoost, Transformer, and Informer. The findings indicate that the developed model demonstrates superior predictive accuracy, thereby offering a reliable foundation for the operational decisions of blast furnace personnel.</p>