This paper investigates a-it application of layered artificial neural network for automatic generation control of the power system. Computer simulations on the interconnected power system with two areas that include reheater effect and also the governor deadband effect show that the artificial neural network control scheme proposed is effective in damping out oscillations resulted by load perturbations. Only one artificial neural network controller, which controls the inputs of each area in the power system together, is considered. By comparing the obtained results with conventional controllers, it is shown that the performance of artificial neural network controller is better than conventional controllers. In this paper, back propagation-through-time algorithm is used as neural network learning rule.