The distribution power systems often supply consumers that contain nonlinear loads; so, the power system operates under a stationary harmonic distorted regime. On the other hand, these consumers have resistive-inductive components that cause the circulation of reactive power and consequently need to use VAr compensation. The interaction between these phenomena can cause parallel resonance events. The research presented in this article had the goal to determine the feasibility of using artificial neural networks to identify parallel resonance in harmonic polluted power systems, more specifically at the consumer's side. The first step in this research was to obtain the necessary data to train, test and validate the proposed methodology. For this, a MATLAB / Simulink model was used. The next step involved the evaluation of the data to determine the correspondence to the parallel resonance events. The research's last step entailed the training and testing of an artificial neuronal network. The results showed that the use of this method is successful in predicting and identifying parallel resonance phenomena in harmonically polluted networks.