This work proposes a thermal model applied to electric vehicle motors based on Newton's Law of Cooling, adding internal heat generation, model predictive control (MPC), and artificial intelligence techniques for nonlinear adjustment of thermal parameters with extension to thermal CFD (Computational Fluid Dynamics). Synthetic data that approximate the operating reality from experimental tests were used for calibration and validation of the model. Loss estimation is included in the model, as well as AI (Artificial Intelligence), with the use of thermal MPC and CFD (Computational Fluid Dynamics). The results indicate that the hybrid approach reduces thermal prediction error and improves the energy efficiency of the cooling system, improving cost in order to maintain the motor temperature within safe operating limits. The contribution is that electric vehicle motors using this thermal modeling system positively predict temperature increases and decreases. (Gisele et al, 2019).