The simulation and prediction of dynamic changes in zooplankton community play a critical role in evaluating the health of bay ecosystems. There remains significant uncertainty in the quantitative prediction of dynamic changes within zooplankton community using classical plankton ecosystem dynamics models. Jiaozhou Bay was used as a representative bay in China’s coastal regions facing multiple pressures from human activities and climate change. Utilizing long-term observation data on zooplankton abundance and environmental factors in Jiaozhou Bay such as temperature, salinity, nutrients, and chlorophyll a concentration from 1997 to 2010, we developed a technical and methodological system for predicting zooplankton abundance in Jiaozhou Bay by applying Empirical Dynamic Modeling approach. Our model achieved an effective prediction with a coefficient of determination (R²) of 0.64, root mean square error (RMSE) of 108.6 and a relative error of 42.5%. After analyzing 15 380 937 potential combinations, we pinpointed temperature and silicate as critical factors influencing the change of zooplankton abundance in Jiaozhou Bay. A deeper look into the research data suggested that the change of zooplankton abundance in Jiaozhou Bay from 1997 to 2010 was a result of interaction between bottom-up control and top-down control in the food web. The quantitative prediction method for changes in zooplankton abundance and the causal identification of driving factors developed in this study exhibited spatial transferability, highlighting their substantial potential as a foundational approach for the future health assessment model of coastal bay ecosystems in China.