Martensite phase and its formation are quite attractive and important in industrial steels for reasons of having good properties such as high strength and high hardness. As such, determining the martensite formation start temperature in steel heat treatment operations is extremely important. Some parameters including chemical composition and grain size are effective factors on this temperature. In this investigation, we have made an attempt to determine this temperature with regard to chemical composition of steels. To reach this goal, we have explored the use of feedforward Artificial Neural Network (ANN) with the Back Propagation (BP) learning algorithm. A comparison is made between the Ms temperatures predicted with this model and those from the empirical equation as well as the experimental values obtained from costly and time-consuming tests in scientific and industrial centers for various steels. This comparison indicates that a better agreement exists between the ANN-predicted results and experimental values than the results from the empirical equation and experimental values.