This paper reports the results of mass loss measurements in the corrosion inhibition of mild steel in different concentrations of H3PO4 in the temperature range 30-60 ℃ using potassium iodide as an inhibitor. The present work is focused on determining the optimum mathematical equation and the ANN architecture in order to gain good prediction properties. Three mathematical equations and three ANN architectures are suggested. Computer aided program was used for developing these models. The results show that the polynomial mathematical equation and multi-layer perception are able to accurately predict the measured data with high correlation coefficients.