This paper implements an artificial neural network (ANN) model able to estimate the internal temperatures in stator winding of a squirrel-cage three-phase induction motor fed by a voltage inverter. A sinusoidal pulse-width modulation (SPWM) inverter feeding three-phase induction motors were tested up to over load. The temperatures are measured in a node of stator winding. The results show that the ANN model produces reliable estimates of interior temperature of stator winding. Through this model, it is possible to determine the temperatures in the tested motors and to analyze the influence of the fundamental frequency voltage on steady state temperatures and motor performance limitation, specifically the maximum acceptable load, so that the motor does not exceed its normal steady state temperature when is supplied by an inverter. Copyright (C) 2010 John Wiley & Sons, Ltd.