<p>Imaging methods offer a computationally efficient alternative to physical property inversion by directly transforming observed data into interpretable spatial distributions for rapid subsurface mapping. However, most existing approaches neglect the influence of rugged terrain on imaging results and suffer from limited resolution. To address these challenges, this study innovatively proposes a focusing-constrained wavenumber-domain imaging method for gravity anomalies in rugged terrain. In this method, a recursive data fitting technique is applied to terrain-distorted observations, generating optimized inputs for imaging calculations. And a terrain-adaptive constraint matrix is developed to mitigate topographic artifacts on the results. Additionally, a focus constraint is implemented during the iterative computation process to enhance spatial resolution. Collectively, these strategies improve both physical accuracy and structural delineation. The method’s effectiveness is validated using two sets of synthetic gravity data simulated with varying terrain complexities and field data from the Sullivan, north Missouri area. Comparative results demonstrate its superiority over conventional 3D iterative imaging under rugged terrain conditions.</p>