Abstract: Objectives: The Gravity Geologic Method (GGM) serves as a key technique for deriving seafloor topography from gravity anomaly data,where the density contrast parameter critically influences inversion results. Conventional GGM methods typically employ a globally optimal density contrast parameter applicable to the entire study area for seafloor topography inversion. However, due to the complex structure and significant spatial variations of seafloor topography in large-scale marine areas, the globally optimal density contrast parameter often fails to accurately characterize the topographic features of localized sea regions. In fact, density contrast parameters vary across different regions, and using a single density contrast parameter leads to reduced inversion performance in the corresponding areas. Methods: This paper proposes a method for calculating optimal local density contrast parameters based on bathymetric zoning, the study area is divided into several subregions using isobaths, and the optimal density contrast parameter for each subregion is derived separately, thereby improving the inversion accuracy of seafloor topography via the gravity-geological method. Taking the central South China Sea (113°E-119°E, 12°N-19°N) as a case study, it is divided into four subregions. Based on SDUST2022GRA gravity data and shipborne bathymetric data from the National Centers for Environmental Information (NCEI), a seafloor topography model (Zone_Model) with a resolution of 1'× 1' is constructed for the study area. Results:Comparison between Zone_Model-predicted depths and ship-measured verification data yielded a standard deviation (STD) of 41.78 m. This represented a 4.95 m accuracy improvement over models employing a single global density contrast parameter. The Zone_Model also surpassed the performance of established global models GEBCO_2024 (STD: 50.86 m) and topo_25.1 (STD: 52.41 m). Analysis revealed that subregions with lower average topographic relief exhibited superior inversion accuracy compared to those with greater relief. Conclusions: The method of dividing the study sea area into subregions according to equal water depth intervals and deriving the optimal density contrast parameter for each subregion is feasible for improving the accuracy of the GGM_Model. When the density of shipborne bathymetric data is sufficient, the smoother the terrain, the higher the accuracy of the inversion results. Topographic relief plays a dominant role in influencing inversion accuracy, while the density of shipborne bathymetric data plays a secondary role.