The objective of this research was to use state-of-the-art artificial neural network approach to estimate the extent and effect of fluctuations in the chemical composition of stainless steel at tapping of an electric arc furnace, and thus scrap and alloy weights in the charge material mix, on the specific electrical energy consumption. Such an estimation would help to further evaluate process control strategies and optimize overall operation of the electric arc furnace. The multilayer perceptron architecture 5-5-1 with hyperbolic tangent function in the hidden layer and linear function in the output layer was used as an optimal neural network model. The model was built, tested and validated based on experimental melts of the electric arc furnace at a melt shop in Italy. The proposed model was presented as an adequate one based on the coefficient of determination (R-2) which was above 0.9 as well as other error parameters calculated. The highest effect on the electrical energy consumption has carbon content. (C) 2015 Elsevier Ltd. All rights reserved.