Pollen analysis, a crucial tool in botany and ecology for examining historical biotic dynamics, has elicited debate owing to its complex link with vegetation. The challenge lies in discerning the ecological significance of pollen data. In this study, we conducted detailed quadrat surveys on Jinhua Mountain, subtropical China, analyzing topsoil pollen to determine whether pollen signals accurately reflect key ecological components in the forests and shrublands. We performed direct comparisons between pollen and plant compositions and calculated pollen percentages and plant Important Values (IVs) for each quadrat. The results indicate greater homogeneity in pollen composition across the study area compared to plant composition, particularly in the high percentage of Pinus pollen. However, distinct plant communities exhibited significantly different pollen compositions, as evidenced by the multi-response permutation test. This divergence aligns with variations in the dominant plant species across different communities. There were significant correlations between pollen percentages and plant IVs, with correlation coefficients of 0.55 (p < 0.001) at the quadrat level and 0.78 (p < 0.001) at the taxon level. These results support the utility of pollen analysis for representing ecologically significant values in subtropical Chinese forests and shrublands. Such correlations might also be extrapolated to pollen-based paleoecological studies.