The Yangtze River Estuary (YRE) serves as a critical land-ocean interface where hydrodynamic mixing and biogeochemical processes jointly regulate dissolved organic matter (DOM) cycling. However, the structural variability of DOM across estuarine gradients remains poorly understood. Here, untargeted metabolomics was applied to investigate the structural diversity and spatial variation of low-molecular-weight DOM (LMW-DOM) in surface waters across the freshwater, transition and adjacent marine zones of the YRE. Among 17,506 quality-filtered molecular features, 1,422 were annotated and classified into 14 superclasses, dominated by organoheterocyclic compounds (24.8%), benzenoids (17.4%), and lipids/lipid-like molecules (16.1%). Multivariate analyses revealed clear compositional differentiation among hydrographic zones, with phosphate identified as the strongest environmental factor associated with LMW-DOM composition. Phosphorus-containing DOM declined significantly from freshwater to the transition zone, suggesting combined influences of sediment resuspension and microbial utilization. Trend-based classification identified four major molecular trajectories across the estuarine gradient: increasing, decreasing, U-shaped, and unimodal patterns. These trajectories were mainly represented by benzenoids, organoheterocyclic compounds, lipids and lipid-like molecules, and phenylpropanoids and polyketides, indicating pronounced restructuring of aromatic, heterocyclic, and lipid-associated compounds during land-to-ocean transport. A subset of these compounds may originate from anthropogenic sources, highlighting complex source contributions and transformation pathways. Co-occurrence network analysis further revealed selective associations between low-abundance bacterial taxa and differential DOM clusters, supporting a potential role of heterotrophic remineralization in shaping molecular patterns. Overall, this study provides new insights into DOM transformation mechanisms at the land-ocean interface and underscores the importance of molecular-level approaches for understanding coastal carbon cycling.