Abstract: The basic method of the neural networks for distinguishing chemical agents were analyzed.For fastness and accuracy,the model of the neural networks distinguishing chemical agents was built by the returning neural networks with deviation unit,including its calculating method and the chemical agents feature extraction based on the wavelet analysis we have analysed the neural networks model for distinguishing chemical agents,put forward the strategy,designed the software,realized the neural networks for distinguishing chemical agents.After analyzing and discussing the results a conclusion is drawn that it is a kind of valid method to realize the automation and intelligence of distinguishing chemical agents using the neural networks combined with chemical sensor.