Thermal stress modeling of power devices is a key factor that influences the design for reliability of Photovoltaic (PV) inverters under long-term operations, i.e., different mission profiles. Due to the requirement of long-term analysis for mapping the inverter reliability more accurately, a thermal model based on a lumped thermal network is normally employed due to its low computational burden. However, there is still a lack of validation in terms of modeling accuracy, e.g., comparing the simulation results against the experimental thermal stress in field operations. Besides, the impact of mission profile dynamics on the accuracy of different thermal modeling approaches have not been analyzed. To address this issue, the model accuracy of two thermal stress modeling approaches for PV inverters are evaluated in this paper by comparing the thermal stress estimated from a thermal model with the experimental results under various mission profile dynamics. According to the results, the average error of the junction temperature estimation is 1.51 % for a transient thermal model and 2.08 % for a steady-state thermal model, respectively. On the other hand, the computational efficiency of the thermal stress modeling can be improved by more than a factor of three when using the steadystate thermal model.