Abstract: The identification model for spontaneous combustion tendency of sulfide ores was established based on BP neural network. This model selected the quantity of oxygen absorption, the self-heating starting temperature and specific surface area of the sulfide ores as the input layer units, and the level of spontaneous combustion tendency as the output layer. Then, it was realized by using MATLAB software, and then applied to the actual identification on the spontaneous combustion tendency of sulfide ores, with correct and reliable predicting results obtained.

Key words: BP neural network, MATLAB, Sulfide ores, Spontaneous combustion tendency