Recently, environmental DNA (eDNA) metabarcoding techniques have been applied to biodiversity investigations in aquatic ecosystems. However, no study has yet tested whether this technique is effective for water bodies in which extensive algal blooms break out. In this study, fish eDNA metabarcoding was carried out in Lake Taihu, which experiences extensive algal blooms, to confirm whether the technique is also effective for fish diversity research in ecosystems with frequent and extensive blooms. In December 2016, three samples were collected, including one collected in the presence of algal blooms and two collected in the absence of algal blooms. In August 2017, six samples were collected, including three collected in the presence of algal blooms and three in the absence of algal blooms. Equal amount of water samples (1 L) was collected from each site; however, the actual amount of filtrate varied with the site. Twenty-seven freshwater fish species were detected from the water samples collected in Lake Taihu. The results showed that the composition of the detected species did not differ whether or not blooms were present. However, the amount of filtration could influence the number of species detected. The results suggest that future eDNA metabarcoding studies under similar water environments should increase the amount of filtration to maximize number of species detected.