Under the background of global climate change and China’s “carbon peak and carbon neutrality” strategy, it is of great significance to assess the carbon sink benefits of urban park plant communities. This study took 20 plant community plots of 20 m × 20 m (400 m²) in Tientsin Water Park as the research object. Carbon sequestration capacity was characterized by carbon stock (CS) and annual carbon sequestration (ACS), and six community structure indicators were quantified: Vegetation Coverage (VC), Canopy Density (CD), Three-Dimensional Green Volume (3DGV), Tree-to-Shrub Ratio (TSR), Vertical Complexity (CV), and Number of Individuals (N). Spearman correlation analysis, principal component analysis, and regression analysis were adopted, and K-means clustering was introduced to identify vegetation structure–function groups, thereby exploring the statistical correlations between these structural characteristic indicators and carbon sink capacity indicators (CS and ACS). The results showed that (1) VC, CD, and 3DGV were significantly positively correlated with CS, suggesting that these factors may be more conducive to long-term carbon pool accumulation; (2) N was significantly positively correlated with ACS, and a nonlinear decreasing trend was observed in the current observation data; (3) the influence of TSR and CV on carbon storage and sequestration also showed a nonlinear correlation. Based on the above correlation findings, the community combinations with higher carbon sink performance in this case were screened out. And suggestions for low-carbon configuration of plant communities, centered on optimizing canopy structure, configuring high-carbon-sequestration tree species, and regulating reasonable density, were proposed, which can be used as a reference for forming hypotheses in subsequent confirmatory studies.