For the reconstruction of the streamwise velocity field in the high‐Reynolds‐number (Re_τ ≈ Ο(10⁶)) near‐neutral atmospheric surface layer (ASL), a predictive model was proposed in our work. In our model, the streamwise velocity time series at each height are predicted by the superposition of mean velocity, small‐scale turbulent signals, and large‐scale turbulent signals. All the parameters used by our model have clear physical meaning. In addition, the parameters can be directly determined by the experiment that is easy to be achieved in the ASL. The model can predict the streamwise velocity between 0.9 and 30 m in the near‐neutral ASL based only on single‐point‐measured large‐scale streamwise velocity signal. For example, mean velocity, turbulence intensity, and power spectrum of streamwise velocity can be predicted by the new model. Furthermore, the streamwise velocity time series predicted by the new model have a good correlation with the directly measured results, especially for large‐scale structures. It indicates that our model can be used to reconstruct the streamwise velocity field in the logarithmic region of the near‐neutral ASL, and our model is an important complement to the Marusic‐Mathis‐Hutchins model.