In order to resolve the problem of deficiency in the data of slab thermal properties during calculating the slab temperature in heating furnace, a new method using artificial neural networks to estimate the slab thermal properties is introduced. Based on the fact of the thermal properties change mainly with the chemical ingredients and the temperature, two BP neural networks which work together and have different function are designed. The experiment study indicates that the prediction precision can meet the engineering requirement. The method decreases the cost, improves the efficiency and supply a new way for insufficient data.