<p>Coastal seas are critical interfaces where sediment transport, hydrodynamic processes, and biological productivity interact. Understanding the spatially heterogeneous drivers of phytoplankton dynamics in such systems is essential for deciphering coastal environmental evolution and mitigating ecological disasters. Using satellite-derived sea surface Chlorophyll-a (Chl-a) concentration data and multi-source environmental data from January 2003 to December 2022, we employed an integrated machine learning approach to investigate the spatiotemporal patterns and environmental controls of sea surface Chl-a concentrations in the waters off the Shandong Peninsula, China. By combining Empirical Orthogonal Function (EOF) analysis, K-means clustering, Random Forest (RF) modeling, and Shapley Additive Explanations (SHAP), we objectively delineated five subregions with distinct seasonal Chl-a patterns. These subregions exhibited contrasting phenology: summer high values occurred in the nearshore bays and the adjacent Bohai Sea (Subregion 1, 4, and 5), whereas autumn-winter and winter-spring high values characterized the eastern coastal current zone (Subregion 2) and the offshore Yellow Sea (Subregion 3), respectively. SHAP analysis revealed that the influence of sea surface temperature (SST), suspended sediment concentration (SSC, a proxy for sediment resuspension and light availability), and mixed layer depth (MLD) varied substantially across subregions—in both magnitude and direction. Furthermore, subregion-specific nonlinear thresholds were identified for these drivers, indicating shifts in dominant limiting factors (e.g., from nutrient to light limitation with increasing SSC). These findings provide quantitative insights into the linkages between sediment dynamics and coastal biogeochemical responses, with implications for understanding eutrophication events, harmful algal blooms, and sedimentary environmental evolution under changing hydroclimatic conditions.</p>