The new technique that combines w ave superposition with the fast Fourier transformation was intro-duced to simulate the nodal three-dimension relevant wind velocity time series of spatialst ructures. The wind ve-locity field where the spatialst ructure is located is assumed to be homogeneous. The wind' spower spectral density is divided into frequency spectral function and coherency function and the spectral functions are transformed as the superposition coefficients. The wavelet analysis has excellent localized characters in both time and frequency do-mains, which not only makes wind velocity time series analysismore accurate, but also can focus on any detail of the objective signal series. The discretew avelet transformation was adopted to decompose and reconstruct the dis-crete wind velocity time series. The stability of wavelet analysis for the wind velocity time series was also proved.