A feeding trial was conducted to determine the optimum level and effect of incremental dietary levels of docosahexaenoic acid (DHA, 22:6n‐3) on growth and non‐specific immune responses in juvenile rock bream, Oplegnathus fasciatus. A basal diet without DHA supplementation was used as a control, and six other diets were prepared by supplementing with 4, 8, 12, 16, 20 or 40 g kg−1 DHA. These diets included no eicosapentaenoic acid and/or arachidonic acid contents. The actual DHA concentrations of the diets were 1, 4.8, 8.9, 13.1, 17.6, 21.2 and 41.4 g kg−1 diet (DHA1.0, DHA4.8, DHA8.9, DHA13.1, DHA17.6, DHA21.2 and DHA41.4 respectively). At the end of feeding trial, final body weight, weight gain, specific growth rate and feed efficiency of fish fed the DHA13.1, DHA17.6, DHA21.2 and DHA41.4 diets were significantly higher than those fed the other diets (P < 0.05). The broken‐line analysis of weight gain indicates that the optimum dietary DHA level is 11.9 g kg−1. Fish fed DHA1.0 had the highest hepatosomatic index, an increase in plasma cholesterol, triglyceride, low‐density lipoprotein and aspartate aminotransferase levels, as well as a decrease in high‐density lipoprotein. Superoxide dismutase activity of fish fed DHA13.1 and DHA17.6 diets was significantly higher than that of fish fed DHA1.0, DHA4.8 and DHA8.9 diets. Fish fed the DHA17.6, DHA21.2 and DHA41.4 diets showed significantly higher lysozyme activity than those fish fed DHA1.0, DHA4.8 and DHA8.9 diets. Therefore, the optimum dietary DHA level could be greater than 11.9 g kg−1 but less than 13.1 g kg−1 in diet.