Abstract: Accurate fitting of coastal wind profiles is of great significance for objective forecasting of winds, fusion of multi-source wind speed observations, and wind energy assessment. Based on observations collected at 32 coastal wind towers, the variation of the power law exponent is analyzed and the impact of different power law exponent quantification methods on vertical wind speed extrapolation is assessed. The results show that the coastal power law exponent varies with observation height, season, and time of day, and the variation trend is spatially different. Further comparative analysis reveals that the power law exponent shows a consistent exponential decay with increasing wind speed. Compared with the fixed value quantification method based on least squares, using the power law exponent that is optimally fitted by an exponential function and fluctuates with wind speed for vertical wind speed extrapolation is meaningful and it can reduce the root mean square error of the extrapolation results by up to about 49%. However, due to the impacts of factors other than wind speed, the fitting accuracy of the power law exponent by wind speed and the effect of its application in fitting the coastal wind profile is unstable. Further research is needed to evaluate the possibility of cutting-edge technologies such as machine learning in the vertical fitting of wind fields.