Abstract: We analyzed the fish assemblage composition and spatio-temporal distribution of fish stocks based on the bottom trawl survey data in Fangchenggang (Qinzhou coastal area of Beibu Gulf) in 2016. Combining with the information of temperature, salinity, dissolved oxygen, pH and depth, we investigated the relationship between fish community structure and environmental factors. A total of 152 fish species belonging to 15 orders, 55 families and 96 genera had been collected, and Parargyrops edita, Secutor Ruconius, Saurida tumbil and Sebastiscus marmoratus were the dominant species. The abundance of fish stocks was the highest in spring. The diversity index was higher in spring and summer, but lower in winter. Four distinct groups were defined by Cluster and non-metric multidimensional scaling (NMDS) analysis, and the average similarity of the groups ranged from 18.27% (Group I) to 32.49% (Group II). The analysis of similarity (ANOSIM) shows that the differences of fish composition among different groups were extremely significant (R=0.703−0.982, P<0.01), with the most sinificant difference of 98.02% between Group I and Group IV. The results of redundancy analysis (RDA) show that the bottom temperature, dissolved oxygen, depth and bottom salinity of bottom were the main environmental factors that affected fish distribution in Fangchenggang.