This study was conducted to compare the effects of manganese sulphate (Mn-S), glycine manganese (Mn-Gly) and manganese 2-hydroxy-4-(methylthio)butyrate (Mn-HMB) on juvenile cobia, Rachycentron canadum L. Treatments consisted of 0, 2, 4, 8, 16 or 32 mg supplemental Mn kg À1 from Mn-S, Mn-Gly or Mn-MHB. Growth performance, manganese status, antioxidant activities and tissue mineral content were analysed after a 70-day feeding period. Specific growth rate (SGR) increased with feeding 6.29 to 12.65 mg Mn kg À1 diet from the Mn-S or 6.86 to 12.39 mg Mn kg À1 from the Mn-Gly or 6.50 to 8.33 mg Mn kg À1 from the Mn-HMB and then plateaued above these levels. Feed conversion ratio (FCR) show decreasing first and then increased trend. Survival rate (SR) were not affected by the dietary treatments (P > 0.05). Fish fed diets supplemented with manganese at levels of 4-32 mg Mn kg À1 had obviously higher hepatic Mn-SOD activity (P < 0.05); on the contrary, hepatic has lower malondialdehyde (MDA) content (P < 0.05) than fish fed the basal diet. The manganese concentrations of whole body and vertebrae increased with increasing dietary Mn levels from 2-32 mg Mn kg À1 (independent on manganese sources). Dietary Mn supplementation did not significantly influence the copper concentrations of whole body and vertebrae, the zinc concentrations of whole body and liver. Analysis by the broken-line regression of SGR indicated that the optimal dietary Mn requirements in juvenile cobia were 15.42, 11.22 and 10.50 mg Mn kg À1 diet from Mn-S, Mn-Gly or Mn-HMB respectively.