Abstract: Ecological connectivity among coastal habitats, including seagrass beds, oyster reefs, and channel wetlands, in the Yellow River Estuary is critical for ecological monitoring and for assessing restoration effectiveness across the national park. In this study, we drew water samples for environmental DNA (eDNA) analysis from four representative sampling stations. Three amplicon markers were employed: MiFish and Tele02 for fish, both targeting the mitochondrial 12S rRNA, and R22mod for eukaryotes associated with the benthos, targeting the 18S rRNA V1—V2 region. Analyses encompassed shared taxa, community differentiation, environmental associations, and cross-marker congruence, integrated with local environmental variables, to evaluate community-level patterns indicative of potential ecological connectivity among the stations. All three markers identified taxa that were common across stations, yet each station retained distinguishable community features, showcasing a blend of cross-station taxon sharing and community differentiation in the Yellow River Estuary. MiFish exhibited stronger station differentiation than Tele02 and revealed a relatively high fish diversity at the oyster reef-associated station S4, where we also observed a greater number of unique fish species and station-specific taxa. Although S4 shared fish species with the other stations, it harbored a comparatively large suite of unique species, underscoring its significance as a priority site for monitoring habitat-specific community dynamics and restoration responses. Beta-diversity partitioning revealed that among-station differences captured by MiFish were predominantly driven by nestedness, indicating that a substantial portion of the observed community differentiation stemmed from variation in species pool size across the stations. In contrast, among-station differences detected by R22mod were primarily attributable to taxon turnover, with R22mod-based community composition showing stronger correlations with salinity and dissolved oxygen gradients, suggesting a more pronounced effect of local environmental filtering on the benthic-associated eukaryotic assemblages. Environmental conditions also varied across the stations: S3 was more aligned with freshwater conditions, S4 displayed the highest salinity and lower dissolved oxygen concentrations, while S1 and S2 occupied an intermediate position. Interestingly, Tele02 showed less among-station differentiation compared to MiFish, yet it yielded a higher proportion of taxa shared across all stations. This implies that the two fish markers captured distinct but complementary dimensions of the fish community structure. Our cross-marker comparison revealed the highest community-structure congruence between MiFish and R22mod, indicating that taxa with divergent ecological traits might jointly respond to the estuarine gradients and local environmental conditions. Overall, multi-marker eDNA analysis enables simultaneous characterization of cross-station taxon sharing, community compositional dissimilarity, and the strength of environmental filtering, thereby revealing a community pattern where potential ecological connectivity and local environmental filtering cooccur within the coastal habitats of the Yellow River Estuary. By integrating complementary information from both fish and benthic-associated eukaryotic taxa, our approach establishes a robust, multi-taxon foundation for long-term monitoring of potential ecological connectivity and for assessing coastal habitat restoration efforts in the Yellow River Estuary National Park Candidate Area.