In recent times, artificial neural networks have shown excellent performance in the area of pattern recognition as an alternative to conventional statistical pattern recognition. The neural network approach assumes no prior knowledge of the underlying distribution of the data and ~djusts to the environment. The back-propagation (BP) model is well-known because it is based on the powerful back propagation training algorithm, which uses the gradient descent method to obtain an optimal least squares solution. It is an unconstrained nonlinear least squares minimization problem which can be solved by an iterative technique. A slow convergence rate and a long training time is expected. Moreover, the network may be trapped in a local minimum, which is a defect of the gradient search method. On the other hand, the radial basis functions (RBF) model uses the linear least squares technique, which guarantees that a solution can be obtained. Unlike the BP model, the weights of the RBF model an~ extracted in one step from the training set. Thus, the RBF training time is less expensive than that of the BP model. In this research, both of the above models are applied to a version of the emitter identification problem. The models are then compared according to the accuracy, generalization capability, network complexity, and effect of noisy data.