Abstract: Juvenile Sebastesschlegelii with body weight of (2.78±0.51)g were set in a vertical recirculating water flume of 100 cm× 25 cm ×25 cm with cube frame reef, cylindrical reef, triangular frame reef and solid cube reef at 6 inflow speeds of 0, 0.1, 0.2, 0.3, 0.4, and 0.5 m/s in order to evaluate the spatial distribution (aggregation) of reef fish in three-dimensional flow field of different artificial reefs by computational fluid dynamics (CFD). The results showed that the maximum upwelling height, upstream area length and side flow area width were not related to inflow velocity. It was found that the oncoming flow area (Area Ⅱ), upwelling area (Area Ⅳ) and side flow area (Area Ⅴ) were not changed with the speed of the inflow velocity. However, as the inflow velocity changes, fish trapping rates were varied with different artificial reefs, with trapping rate being reduced from 63.06% to 52.50% in the cube frame reef, being increased from 43.06% to 59.44% in the cylindrical reef, increased from 39.17% to 83.61% in the triangular frame reef, and increased from 49.72% to 70.28% in the solid cube reef. The significantly high aggregation of fish was observed in high-velocity environments, and the descending order of the frequency of occurrence of distribution quantity in five areas of reefs was expressed as the back eddy area (Area Ⅲ) > the side flow area (Area Ⅴ) > the reef inner area (Area Ⅰ) > the oncoming flow area (Area Ⅱ) > the upwelling area (Area Ⅳ). The findings indicated that the triangular frame reef with complex internal structure had the best trapping effect on experimental fish among the four types of artificial reefs.