Abstract: The prediction models of carbon steel and low alloy steel in atmosphere between corrosion rate exposure time were achieved,according to environment data and material corrosion data collected by atmospheric corrosion sites of China and using artificial neural network method.Relationship between corrosion rate and exposure times for carbon steel and low alloy steel in atmosphere was stuelied chieved, a based on environment data and material corrosion data and using studied artifial neural network methed. And the prediction models for predicting corrosion value of metal in atmosphere were achieved.