Coral reef sh populations are declining due to multiple factors including habitat destruction, over shing, and climate change. While these can include seasonal impact, seasonal changes can also be a confounding factor, so that seasonal monitoring is essential for detecting true temporal changes in sh community abundance and composition. However, seasonal monitoring is rarely implemented. The aim of this study was to detect the seasonal patterns of coral reef sh community structure around Tidung Kecil Island, Indonesia, using eDNA metabarcoding (eDNA) and underwater visual census (UVC). The UVC identi ed 32 species belonging to 10 families in the rainy season and 29 species belonging to 7 families in the dry seasons. The eDNA metabarcoding identi ed 209 species belonging to 56 families and 27 species belonging to 17 families in the rainy and dry seasons, respectively. Based on eDNA metabarcoding data, coral reef sh abundance and community composition differed signi cantly between seasons (Mann Whitney, p<0.01), while the UVC method did not detect these seasonal differences (Mann Whitney, p>0.05). UVC and eDNA data both showed a predominance of omnivorous shes in the rainy season and carnivorous shes in the dry season. Ecological indices did not differ signi cantly between seasons for either method, although the species making the highest contributions to the similarity (SIMPER) analysis differed between methods. Overall, this study con rms the premise that eDNA metabarcoding can be an effective tool for monitoring seasonal variation in coral reef sh communities.