Determination of percentage of protein in a pork carcass is a very costly and lengthy process. Therefore, we attempted to identify an easy and economical means to predict protein content of pork carcasses. Forty-seven market-weight pigs were slaughtered. The right side of each carcass was skinned, deboned and ground for composition analysis. Carcass data were collected on the left side and specific gravity analysis was performed on the side, ham and 8-9-10 rib loin section of each carcass. The hams and loin sections were dissected into skin, bone, subcutaneous fat, seam fat and individual muscles. After these components were weighted, they were analyzed for chemical composition. Percentage of chemical protein of each animal was determined and correlated with each of the weights and measurements taken on the carcass, ham and loin sections. Stepwise regression was used to predict decimal fraction of protein using either the carcass data, specific gravity ham measurements, loin section measurements or various combinations of these. In prediction equation, specific gravity of the carcass accounted for 73.4% of the variation found in percentage of protein. Prediction equations using indicators from the ham and loin section resulted in equations with R-square values as high as .783. These results suggest that specific gravity or indicators from the carcass, ham or loin section can be employed to predict percentage of protein in pork carcasses.