Abstract: Based on environmental DNA (eDNA) metabarcoding technology, we investigated the fish biodiversity of the Hengqin National Wetland Park using, so as to assess the current status of fish diversity and evaluate the effectiveness of conservation and management efforts. High-throughput sequencing detected a total of 55 fish species, belonging to 8 orders, 23 families, and 45 genera. Perciformes was the dominant order, with a relatively high proportion of omnivorous and carnivorous species. Spatially, the restoration area exhibited the highest species richness, while the Tianmu River had the lowest diversity. In terms of species dominance, five species, including Ctenopharyngodon idella, Oreochromis niloticus, Sarotherodon galilaeus, Coptodon zillii and Cyprinus carpio, were identified as dominant taxa. Principal Coordinates Analysis (PCoA) reveals that District 1 was most similar in community composition to Mangzhou, the fish ponds and the restoration area, whereas the fish community in the Tianmu River was markedly distinct from the other four sites. These findings highlight the significant potential of eDNA technology in rapid biodiversity assessment and ecological pattern analysis. Based on the results, we propose the following recommendations for local aquatic biodiversity conservation: 1) conduct systematic baseline surveys of biological resources within the wetland park; 2) establish a dynamic monitoring system for invasive alien fish species; 3) enhance environmental monitoring and assessment.