Abstract: There are many factors that can affect the representation of wind tower data. These factors should be completely considered in order to obtain the reasonable fetch of wind tower data. Footprint function, usually applied to investigate the representation of turbulent flux, is a function of the wind speed, stability, and overlying surface roughness length, etc. Footprint function can be used to quantitatively characterize the representation of observational data at one spot. Thereby, it has been introduced to study the representation of wind tower data collected at 44 wind towers on various overlying surfaces in 2010. These wind towers are part of the national wind energy resource observation net. The stability parameter Lp is calculated first using the observed wind speeds and temperature at two levels (10 and 70 m) and data under the condition of neutral stratification is determined based on Lp. The roughness length is then calculated using the data that follow the logarithm wind profile in the neutral stratification condition. The data in prevailing and sub-prevailing wind directions at 70 m under neutral stratification condition is applied to investigate the applicability of two footprint functions (Schuepp and Hsieh functions) and the relationship between the roughness length and the fetch of wind tower data. The fetches obtained by the two methods are close to each other. However, Schuepp method cannot be used in complex terrain area. The relationship between the fetch and the roughness length can be described by a log-linear function. The distance at 90% of integral footprint function and the location at which the overlying surface changes obviously are recommended as two critical parameters to determine the representative distance of wind tower data in this study. However, trajectory experiments or numerical simulations are still needed for the determination of the threshold value of footprint function.