Abstract: An artificial neural network was established to forecast the surface tension of pure metal from the experimental data of atomic radius, melting point, boiling point and atomization enthalpies. The trained network can represent the relahonship between the input factors and output factor (surface tension).The associated and forecast data for more than 40 pure metals are acceptable considering the deviation of the experimental dara for surface tension, which shows a good prospect of artificial neural network in the prediction of surface tension of pure metals.