Freshwater bivalves contribute to key ecological functions in lake ecosystems, yet their cryptic and benthic lifestyles often hinder detection through conventional surveys. In this study, we applied environmental DNA (eDNA) metabarcoding to assess the diversity and distribution of unionid bivalves in six lakes across Republic of Korea. Water samples were collected from three sampling strategies-Center Surface, Center Mix, and Waterside Surface-and processed using 16S rDNA-targeted primers followed by high-throughput sequencing. A total of four unionid species (Cristaria plicata, Sinanodonta lauta, Unio (Nodularia) douglasiae, and Anodonta woodiana) were detected across 18 sampling points. Notably, eDNA successfully identified unionid presence in all lakes, even where conventional surveys failed to observe individuals. Among the sampling strategies, Center Mix exhibited the highest values for Shannon and Simpson indices as well as ASV richness. Waterside Surface samples generally showed lower diversity and detection frequency. A Venn diagram of ASV occurrences revealed three ASVs shared across all sampling strategies and one unique ASV found only in Center Mix. These results indicate that sampling location significantly affects detection sensitivity and diversity representation in eDNA-based bivalve monitoring. Combined application of Center Mix and Center Surface strategies may enhance both detection efficiency and species diversity coverage in lentic environments.