Abstract: Artificial Neural network (ANN) is a new approach for data assimilation process. The performance of two feedforward (multilayer perceptron and radial basis function), and two recurrent (Elman and Jordan) ANNs is analized. The Lorenz system under chaotic regime is used as a test problem. These four NNs were trainned for emulating a Kalman filter using cross validation scheme. Multilayer perceptron and Elman ANNs show better results. The results obtained encouraging the application of the ANNs as an assimilation tecnique.