A proper and stable accretion layer on lining could prolong campaign life of blast furnace and promote energy efficiency. In order to avoid forming abnormal scaffold on shaft wall, an estimated model of lining residual thickness and accretion thickness was established on the base of the temperature and pressure data of blast furnace shaft. And, the distribution pattern of gas flow was classified by using a neural network based on the gas temperature measurements from the above-burden probes. It could monitor the lining residual thickness and accretion thickness, the trend of heat load on wall, and gas flow distribution. According to the gas temperature distribution map, the relationship between the heat load and several process variables were established to provide a useful tool for operating guidance in daily practice. As any abnormalities occurred, operator can take countermeasure in time to keep the furnace in stable running.