Characterization of the power semiconductor device switching process is critical to application since the switching speed determines the performance of the power converter. Accurate characterization of silicon carbide (SiC) devices is more difficult than the silicon counterparts due to the higher switching slew rate. It can lead to ringing in the power loop and finally the errors in characterization. Aiming at addressing this problem, in this paper, an accurate mathematical model of switching trajectory is derived to extract the accurate voltage and energy loss during SiC MOSFET switching process. A model-based true waveform restoration method is proposed to mitigate the impact of parasitic parameters on dynamic characterization. Without changing the hardware, it can increase the accuracy of the dynamic analysis of the SiC devices.