Abstract: This study establishes an eco-hydrodynamic model for the coastal waters off Hainan Island using the coupled FVCOM-FABM-ERSEM framework, aiming to systematically investigate the process of red tide outbreaks. To enhance the diversity of phytoplankton growth dynamics in the ERSEM module, a multiplicative interaction model is introduced. In addition, sediment-induced light attenuation and salinity inhibition are incorporated to improve the accuracy of simulations in estuarine transition zones. Model results show good agreement with observations, successfully reproducing the distribution patterns of nutrients and phytoplankton communities. The study reveals the full progression of a red tide event: initially, nutrient input from the Nandu River, strong turbulence, and low temperatures limit phytoplankton uptake, maintaining a high-nutrient background; subsequently, a rapid rise in water temperature lifts growth inhibition, triggering a bloom of Phaeocystis globosa, which quickly becomes the dominant species; eventually, single cells aggregate into colonies, leading to the formation of a red tide. Correlation analysis indicates that water temperature is the primary driver of the outbreak. With parameter adjustments, the model can be applied to a wider range of marine environments, providing a scientific basis for red tide forecasting and early warning systems.