High-precision residual time picking within Common Imaging Gather is regarded as the fundamental basis of velocity inversion, and stable input data are supplied for the inversion process. A residual moveout estimation approach defined in the Cartesian coordinate system is proposed in this paper according to the inherent features of conventional CIG gathers. In this Cartesian coordinate framework, the depth residual moveout is characterized as a curve on a two-dimensional section, with its relationship to the slope vector described by the Poisson equation. By solving the Poisson equation using a plane-wave destructor algorithm, vector decomposition is performed on the migrated CIG gathers in the spatial domain to derive the residual moveout. The picked residual moveout is then utilized in grid tomography for velocity model building to obtain the optimal imaging velocity. The insufficient imaging precision induced by velocity errors is effectively resolved with the proposed method, and favorable imaging outcomes are achieved.