This review examines advancements in potential functions for hydrate⁃bearing sediments, analyzing the evolution of single⁃phase optimization, multiphase interface parameter correction, and machine learning integration methodologies to elucidate how parameterization strategies regulate multiphysics simulation accuracy. Current potential functions face challenges in extreme temperature⁃pressure adaptability,multiphase parameter compatibility,computational efficiency⁃accuracy balance,and experimental validation. Future research should focus on developing extreme⁃environment adaptive multiphase potential function systems, constructing multi⁃physics coupling models integrating traditional force fields with machine learning potentials, and establishing experimental⁃simulation collaborative validation frameworks. to provide multi⁃scale theoretical support from atomic mechanisms to reservoir responses for safe hydrate exploitation.