Abstract: In recent years, efficient electrical equipment for reducing energy consumption has drawn increasing worldwide attention. Although silicon (Si) has been used as a power semiconductor device, its improving effect on the performance of power semiconductor devices is greatly limited by its physical characteristics. Compared with Si, silicon carbide (SiC) as a type of wideband gap semiconductor has more excellent comprehensive physical properties in power device applications, including a triple wideband gap, a triple high thermal conductivity, and a tenfold breakdown electric field. Moreover, SiC can form silicon dioxide (SiO₂) on the surface through thermal oxidation, which plays an important role in device manufacturing technology as an insulating layer. Based on these properties, SiC has gradually replaced Si as the preferred material of power devices used in metal oxide field-effect transistors (MOSFETs). The structure of a MOSFET contains a polysilicon-oxide layer (mostly SiO₂)-SiC or diamond as the core. This structure is exactly equivalent to that of a capacitor, with SiO₂ as the dielectric medium in the middle, and the capacitance value is determined by the thickness and dielectric coefficient of SiO₂. However, the anisotropic process during the thermal oxidation from SiC to SiO₂ results in a large difference in oxidation rate on different crystal faces, which adversely affects the performance of semiconductor devices. Therefore, studying the growth law of SiO₂ on each crystal surface of SiC is of vital importance. Effective and reasonable dynamic models are expected to clarify the behavior. In this paper, the representative modified Deal-Grove model (Song model and Massoud empirical relation) and Si−C emission model were researched and compared systematically in terms of the reaction mechanism and fitting accuracy. On this basis, the advantages and disadvantages of the models were analyzed, and the possibility of the application of the real physical picture model established by our research group was proposed, which can further contribute to optimization and modification for the precise description of the oxidation kinetics of SiC on different crystal faces.