Abstract: To establish expressions for determining high-level wind profiles using ground-based observations, this research validated various methods to determine the roughness, boundary layer height, and geostrophic wind speed, based on 1-year of observations from the Xilinhot National Climatological Observatory. On this basis, the average wind profiles under different stability conditions were compared with the caculation results of established boundary layer wind profile models. The results demonstrated that the predicted geostrophic wind speeds based on geostrophic resistance drag parameters obtained from reanalysis data are close to the actual wind speeds. The logarithmic wind profile model can provide accuracy and simple cauculation under neutral and unstable conditions. The Deaves–Harris（DH） model was suitable for strong winds. The modified mixing-length Gryning model and the two-layer wind profile model that considers the Ekman layer Coriolis force can obtain relative accurate simulations under different stability conditions, albeit with the disadvantage of requiring complex computational procedures. The improved Gryning model is suitable for simulating wind profiles for the nocturnal residual layer.