Abstract: The turbulence movement can be treated as consisting of many eddies of different time-spatial scales, and spectrum analysis can separate these eddies. On the basis of the introduction to the Hilbert-Huang transform (HHT), HHT is applied into a series of the turbulence data in the Horqin Sandy Land during the periods of 00:00－01:00, 05:00－06:00 and 12:00－13:00 BT on 16 April 2011. It is shown that the central instantaneous frequencies of Intrinsic Mode Functions (IMFs) are able to show the turbulence fluctuations at the different scales which include the gust in 3-5 min and the internal gravity wave in 10－13 min as in the Hilbert marginal spectrums. The 3D Hilbert spectrums denote that turbulence energy concentrates in the bands lower than 0.5 Hz. Corresponding with the strong turbulence periods, IMFs and the Hilbert spectrums contain more power as well. Compared with the traditional technique (Fast Fourier Transform, FFT), HHT can perform the distribution of spectrum more clearly.