The uncertainty of parameterization schemes makes it difficult for numerical simulations to describe atmospheric boundary layer processes accurately, and has, therefore, been the focus of many researchers in recent years. Four boundary layer schemes, namely the WRF model YSU, ACM2, QNSE, and BouLac, were used to conduct numerical simulation experiments on the atmospheric boundary layer in winter in Southeast Tibet. Radio sounding observations from January 3 to January 9, 2022, were used for validating the atmospheric boundary layer structural characteristics, including temperature, specific humidity, wind direction, wind speed, as well as the modeled results for the near-surface stratum, surface temperature, and heat fluxes. Subsequently, the applicability of different boundary layer parameterization schemes in Nyingchi, Southeast Tibet, was evaluated. Results show that the ACM2 scheme exhibits the smallest simulation deviation for the potential temperature. When convective exchange is weak, the parameterization scheme has a small boundary layer simulation error. Local versus nonlocal mixing contributes more to the boundary layer development than turbulent kinetic energy. For the boundary layer height, the effect of the TKE scheme is greater than that of the nonlocal scheme. For specific humidity, the simulations show significant drying out, and the BouLac scenario is overall the closest to the observations. For wind speed, simulations are more consistent with the observations. For the surface air temperature and surface temperature, the simulated values of the parameterized schemes are more consistent with the trend of the observed values, and the ACM2 scheme is the most effective. In winter, latent heat flux is low, sensible heat flux plays a dominant role, and the BouLac scheme simulates them most appropriately.