In recent years, thermoplastic composites have found their place in large business sectors and are in direct rivalry to thermoset matrix composites. In order to ensure efficient and lean processes, process modeling gains ever-growing attention. This work shows the computational fluid dynamics (CFD)-modeling of a typical heating step in a thermoforming process of a thermoplastic composite sheet. When heating thermoplastic composites, the heat conduction proceeds anisotropic, and the sheets are subject to thermal deconsolidation when heated above the melting temperature of the polymer matrix adding to the anisotropic effect. These effects are neglected in known process models and this study shows the first successful attempt at introducing them into CFD-modeling of the heating of thermoplastic composite sheets. Thus, the simulation requires temperature dependent values for the anisotropic thermal conductivity and the coefficient of linear thermal expansion, which are calculated with novel physical models which were developed solely for this cause. This alters the behavior of an isotropic CFD-model and allows the successful validation via laboratory experiments using glass fiber reinforced polypropylene (PP/GF) sheets with embedded thermocouples to check the internal temperature distribution when the sheet is heated to the designated forming temperature in a composite thermoforming press. The incorporation of this newly developed process model reduces the error in the core temperature prediction from close to 70 °C to 3 °C at the forming temperature.