Starving and re-feeding trials were conducted to evaluate the effects of starvation duration and recovery times on the growth performance and physiological characteristics of juvenile tongue sole Cynoglossus semilaevis, which included growth, body composition, intestinal morphology and digestive enzymes activities. The fish starved for one(D1), two(D2) and four(D4) days, respectively, were recovered for three(R3, R6, R12), seven(R7, R14, R28) and eleven(R11, R22, R44) folds of starvation days while the control fish were fed constantly. The experiment was conducted for 96 days. The fish individuals in D1R11, D2R14 and D2R22 caught up in weight with the control, indicating that complete compensatory growth existed in these fish individuals. The highest body weight gain and total specific growth rates were observed in D2R22. Food intake increased significantly in most treatments in comparison with that of the control except for D1R7 and D2R22(P < 0.05). Food conversion efficiency and apparent digest coefficient in D2R22 were significantly promoted, being higher than that of the control(P < 0.05). The activities of trypsin and lipase were found to be closely related to the growth performance. The activities of liver and hindgut trypsin, also midgut and hindgut lipase in D2R22 were significantly higher than those of the control(P < 0.05). The fold height(HF) of foregut and midgut increased significantly in D2R22, and HF decreased significantly in D1R3, D2R6 and D4R12(P < 0.05). On our findings, we may conclude that the optimum starving and re-feeding strategy is starving for 2 days and re-feeding for 22 days. And in this starving and re-feeding strategy, the compensatory growth could be mostly attributed to the promotion of food conversion efficiency and digestibility coefficiency. Starving and re-feeding trials were conducted to evaluate the effects of starvation duration and recovery times on the growth performance and physiological characteristics of juvenile tongue sole Cynoglossus semilaevis, which included growth, body composition, intestinal morphology and digestive enzymes activities. seven (R7, R14, R28) and eleven (R11, R22, R44) folds of starvation (R11, R22, R44) days while the control fish were fed constantly. The experiment was conducted for 96 days. The fish individuals in D1R11, D2R14 and D2R22 caught up in weight with the control, indicating that complete compensatory growth existed in these fish individuals. The highest body weight gain Food intake increased significantly in most treatments in comparison with that of the control except for D1R7 and D2R22 (P <0.05). Food conversion efficiency and appa rent activities digest coefficient in D2R22 were significantly promoted, being higher than that of the control (P <0.05). The activities of trypsin and lipase were found to be closely related to the growth performance. The activities of liver and hindgut trypsin, also midgut and The fold height (HF) of foregut and midgut increased significantly in D2R22, and HF decreased significantly in D1R3, D2R6 and D4R12 (P <0.05). On our findings, we may conclude that the optimum starving and re-feeding strategy is starving for 2 days and re-feeding for 22 days. And in this starving and re-feeding strategy, the compensatory growth could be mostly attributed to the promotion of food conversion efficiency and digestibility coefficiency