This paper presents an application of layered artificial neural network controller (ANN) to study automatic generation control (AGC) problem in a four-area interconnected power system that three areas include steam turbines and the other area includes a hydro turbine. Each area of steam turbine in the system contains the reheat effect non-linearity of the steam turbine and the area of hydro turbine contains upper and lower constraints for generation rate. Only one ANN controller, which controls the inputs of each area in the power system together, is considered. In the study, back propagation-through-time algorithm is used as ANN learning rule. By comparing the results for both cases, the performance of ANN controller is better than conventional controllers. (C) 2002 Elsevier Science Ltd. All rights reserved.