The ecological evolution of the Changjiang River Estuary, influenced by river discharge, ocean currents, and anthropogenic activities, has undergone significant ecological changes in recent decades. Metabarcoding was used to analyze sedimentary ancient DNA (SedaDNA) from a sediment core (Sed.C4-3) collected from the estuary in China. By integrating high-throughput sequencing and multivariate statistical analyses, we were able to reconstruct plankton community dynamics over the past ~75 years. Results show that the planktonic community underwent two major turning points, around 1980 and 2010. The communities dominated by large zooplankton and fungi were replaced by microplankton-dominated assemblages, particularly dinoflagellates and ciliates. This shift reflects a transformation from a system with low species turnover and weak connectivity to one with high modularity and strong interlinkages. Beta diversity partitioning revealed that community heterogeneity was mainly driven by species turnover rather than nestedness. Turnover rates have decreased since 2010, showing greater structural convergence among communities. Through co-occurrence network analysis, we found evolving successional paths and shifting ecological roles of keystone species over time, indicating that increased system stability may come with hidden vulnerabilities. This research demonstrates the importance of sedaDNA in revealing the historical ecological dynamics of marine ecosystems, providing valuable information on how coastal biodiversity responds to environmental changes.