An experimental feeding trial was conducted to determine the optimum level of dietary protein on growth performance, feed utilization, survival and carcass composition of yellowfin seabream, Acanthopagrus arabicus. Five semi-purified experimental diets were formulated containing 350 (P30), 400 (P35), 450 (P40), 500 (P45) and 550 (P50) grams of protein kg-1 of dry matter. Thirty healthy fish (20.94±0.81g initial weight) were stocked in each floating net cage (1.5m x 1.5m x 1.5m), with three replicates of each treatment. The diets were fed to visual apparent satiation two times a day for a period of eight weeks. At the end of the trial, P40 appeared to have significant (P<0.05) influence on the growth performance. The percent weight gain, average daily weight gain and specific growth rate (SGR) increased significantly by increasing dietary protein level from P30 to P40 (P < 0.05), and further increases in protein level (P45, P50) reduced growth. The best feed conversion ratio (FCR; 1.80±0.03) was also noted at P40 protein level. No significant (P>0.05) differences were found in the whole fish body proximate composition of protein, ash and moisture fed at different protein levels except for lipid content of fish body influenced significantly (P<0.05) by increasing protein in the diet. The second degree polynomial regression estimates that A. arabicus grows best when fed on a diet containing up to 42% protein in floating net cages.