This work presents the development and validation of a model for electric networks supplying nonlinear loads (NLL). The model is based on Neural Networks (NN) and its purpose is the prediction of harmonic currents sank by the loads, in case of supply impedance changes caused by the insertion or releasing of neighboring loads or by the voluntary insertion of filters to reduce such harmonic currents. The NN is trained using data obtained from several Matlab simulations and the model is validated using the same network with different supply impedances and load conditions.