Abstract: Aiming at the complexity of coke oven heating and combustion process,a information-fusion-based method is proposed to judge the operating states of coke oven heating and combustion process.At first,data correlation is analyzed,and consistency and accuracy of the sensed data used to analyze the operating states are judged.Then a two layer information-fusion method is used to judge the operating states.According to the analysis of the features of raw gas temperature,data of the feature points of each sensor is extracted and the method of expert rules is used to analyze the operating states in every coking chamber.The judgment results of each coking chamber are arranged and classified,and a fuzzy-neural network method is used to make the second information fusion and to get the realtime operating states of the coke oven heating and combustion process.This method has been applied to real world,and its validity is verified.