Modern electrical systems have a significant presence of electronic loads, which in turn produce negative impacts on distribution systems and loads, this has motivated their study to be increasingly prioritized, aiming to reduce their impacts from corrective actions. Harmonics are classified as positive, negative, and zero sequence, and their impacts on loads can vary according to the harmonic present. In the case of electric motors, negative sequence harmonics result in the greatest impacts. This work presents a classifier of existing harmonics in the input waveform of electric motors classes IE2, IE3 and IE4 using artificial neural networks (ANN), for that purpose, negative (2nd), positive (7th) and zero sequence harmonics (3rd) were applied separately and combined in the electric motors, the data was exported for a classification algorithm to identify existing harmonics. The results show how the algorithm presents good approximations of the present harmonics, mainly with those of positive and negative sequence.