As one of the important links during the traditional long - process iron and steel production, sintering mineral and quality directly affect the efficiency of ironmaking production, and the sintering endpoint is taken as an important indicator to detect the production status of sintering, which will have an effective role in promoting production and accurate prediction. A deep neural network (DNN) prediction model based on Python language is designed, and the trained model is applied to the sintering endpoint prediction, and the mean square deviation loss of the final model reaches 0.375, which is highly accurate and can meet the field requirements. At the same time, starting from the important parameters affecting the sintering endpoint, a fuzzy controller at the sintering endpoint is constructed to adjust the sintering endpoint, and an intelligent adjustment model for ignition temperature, an optimization model for mixed moisture and an interval optimization control method for chemical composition of sinter are constructed to assist in adjusting the sintering endpoint. Finally, the development of the sintering endpoint prediction and operation guidance system is completed in Python language, the accurate prediction of the sintering endpoint and the all - round and multi - angle intelligent processing after the sintering endpoint anomaly are realized, and an intelligent control system integrating online monitoring, endpoint prediction, endpoint prediction model optimization, abnormal alarm and exception handling into one is constructed, which provides a favorable guidance for sintering production personnel to realize the optimization of the whole process of sintering endpoint.