A 12-week feeding trial was conducted to determine the optimal dietary selenium requirement for coho salmon (Oncorhynchus kisutch W.) alevins. Six experimental diets (contained about 42.2% crude protein and 10.5% crude lipid) were formulated to contain selenium level of 0.07, 0.15, 0.24, 0.41, 0.75, and 1.38 mg/kg, respectively. Each diet was fed to triplicate groups of fish with initial mean body weight of 0.38 +/- 0.01 g, which were randomly assigned to 18 plastic tanks (80 x 60 x 60 cm, length x width x height, water volume 240 L), and each tank was stocked initially with 50 fish. Fish were fed four times daily and used the method of overfeeding (15-20% body weight) to fish. The sampling procedure was implemented after 24 h of fasting. Results showed that specific growth rate (SGR) in 0.24, 0.41, and 0.75 mg Se/kg groups were significantly higher than those in 0.07 and 1.38 mg Se/kg groups (P < .05). There were no significant differences in hepatosomatic index (HIS), intestinal somatic index (ISI), condition factor (CF), and whole-body composition among dietary groups (P > .05). Hepatic malondialdehyde (MDA) content significantly declined with increasing dietary up to 0.41 mg Se/kg and increased as selenium further increased. The hepatic superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GPx) significantly increased with increasing dietary selenium up to 0.41 mg/kg and decreased significantly as selenium further increased (P < .05). Fish fed diet with 0.75 mg Se/kg had the highest glutathione reductase (GR), which did not differ significantly with 0.41 and 1.38 mg Se/kg groups (P > .05) but significantly higher than those in 0.07, 0.15, and 0.24 mg Se/kg groups (P < .05). Broken-line regression analyses based on SGR, hepatic, and whole-body selenium contents indicated that the optimum dietary selenium requirements were in the range of 0.39-0.43 mg/kg for coho salmon alevins.