SUMMARYDuring early stage of primary restoration process, unexpected overvoltage may happen due to nonlinear interaction between the unloaded transformer and the transmission system. Such an overvoltage might damage some equipment and delay power system restoration. Artificial neural network (ANN) is addressed in this work, in order to estimate the temporary overvoltages (TOVs) due to transformer energization. In the proposed methodology, Levenberg-Marquardt second order method is used to train the multilayer perceptron. Equivalent parameters of the network are added to ANN inputs to achieve good generalization capability for trained ANN. The developed ANN is trained with the worst case scenario of switching angle and remanent flux, and tested for typical cases. The simulated results for a partial of 39-bus New England test system, show that the proposed technique can estimate the peak values and duration of switching overvoltages with good accuracy.