In practice, achieving a balance between accuracy, stability, and computational efficiency in modeling contaminant dispersion in marine environments remains challenging due to complex physical dynamics and numerical constraints. To address these challenges, an enhanced cell-based model (CBM) is developed and applied to simulate pollutant transport in the ocean. The CBM discretizes the spatial domain into uniform cells, resulting in a naturally parallelizable structure, and characterizes the transport process by incorporating both water flow-driven convection and diffusion effects. Moreover, two approaches are proposed for estimating the diffusion coefficient, and their performance is compared to a first-order upwind scheme finite-difference method (FDM) solution. Finally, the CBM is comprehensively compared with both the FDM and the finite-element method (FEM) solvers under varying spatial and temporal resolutions. Simulation results show that the CBM is less affected by the Courant-Friedrichs-Lewy (CFL) conditions and demonstrates stable convergence where the FDM fails or requires stricter settings. In addition, the CBM offers a favorable trade-off between accuracy and computational efficiency under coarse configurations. These results indicate that the CBM provides a reliable foundation for dynamic modeling and integration with learning-based frameworks in marine environment simulations.