Density imaging is a method of inverting the sub-surface density distribution according to the spectrum of the gravity and gravity gradient in the wavenumber domain. This method effectively gives full play to the characteristics of fast calculation in the wavenumber domain, improves the computation efficiency, and creates an accurate 3D sub-surface density model. In this paper, the corresponding relation between the gravity and gravity gradient anomalies and the model, and their spectral characteristics were analyzed, which according to preliminary inverse. Then, the 3D density imaging of gravity and gravity gradient was performed on the theoretical data and its noise-added data in the wavenumber domain with depth weighing, and a density model consistent with the theoretical model was obtained. The strong anti-noise capacity of the density imaging method was proved. Finally, the method was verified in the Decorah area of the United States, and the characteristics of gravity and gravity gradient anomalies measured in this area were analyzed, and the 3D density imaging of gravity and gravity gradient was performed in the wavenumber domain. The location of the siliceous intrusive rocks with the relatively low-density and the Decorah complex with the relatively high-density, and the intrusive rock mass with the relatively highdensity distributed in the surrounding rock were obtained through inversion. A clear understanding of the intrusive pathways to the rock mass was obtained, and the effectiveness of the density imaging method has been verified. This provides support for further understanding of the structural division and geological evolution in this area.