Abstract: The study on full-scale wind field characteristics is one of the fundamental works in the structural wind-resistance design. Being one of the main parameters among them, the wind speed spectrum describes the distribution of turbulence energy with the frequency. Van der Hoven first proposed the wind speed spectrum in the wide frequency range, including the macro-meteorological and micro-meteorological peaks, which made a pioneering contribution in wind engineering. Considering that the data acquisition equipment and computing ability in early years were relatively poor and the urbanization is accelerating recently, the full-scale wind speed spectrum should be further examined in urban areas. The paper obtains the wind speed spectrum in a wide frequency range including the macro-meteorological and micro-meteorological peaks based on the wind speed data collected by the sonic anemometers at the heights of 47 m, 80 m and 140 m of the Beijing Meteorological Tower in 2013. The wind speed spectrum is compared with the Van der Hoven spectrum and other previous spectra. The results show that for the macro-meteorological spectrum, besides around the 4-day cycle, the periodic deterministic component with a period of 24 hours is also detectable and prominent. However, the spectrum does not show significant peaks at a period of 12 hours. The micro-meteorological peak is relatively weak compared with the macro-meteorological peak. The macro-meteorological peak increases with the height, while the micro-meteorological peak shows little change for different heights. Considering that the wind spectrum in the high frequency range is the main focus of the structural wind-resistance design, the micro-meteorological spectrum is recalculated based on the strong wind speed samples whose mean wind speeds are larger than 8 m/s. The comparisons with typical power spectrum are performed. It is shown that the micro-meteorological spectrum for different heights generally correspond with each other, and they are shown to be close to the Kaimal spectrum.