Artificial neural network (ANN) based controller for the excitation system of electrical alternator (synchronous generator) is presented in this paper. Feed forward neural networks architecture of ANNs (Artificial Neural Networks) has enjoyed many successful applications in the field of power systems and control. These networks include two major categories: RBF (Radial Basis Function) network and MLP (Multiplayer Perceptrons). This paper explores the suitability and applicability of RBF architecture of ANN’s in the excitation system of synchronous generator. Computer based simulation of a synchronous generator subject to different operating condition is performed to demonstrate the effectiveness of the proposed excitation system. The performance of the trained neural networks is compared with PID (Proportional Integral Derivative) excitation system of synchronous generator and it is found that the RBF architecture of ANNs can show not only a good scope of application but they are also suitable for various operating conditions of synchronous generator.