A 4 × 2 factorial experiment was conducted to investigate the effect of a processed soy protein (PSP) on broiler performance and digestive physiology. Four inclusion levels of PSP (0, 25, 50 or 100 g/kg in either corn or wheat-based diets) were used in the starter diets. Feed intake was significantly lower (P < 0.01) on corn-based diets than those on wheat-based diets at 35 d. Over the first 10 d, chicks on corn-based diets tended (P < 0.09) to have higher body weight (BW) than wheat-based diet chicks. Across the 35 d trial, PSP level showed a strong tendency (P < 0.06) to be related to higher BW in birds, regardless of grain type. Both grain and PSP experimental factors significantly (P < 0.01, and P < 0.001) interacted at an early age, improving BW and feed conversion ratio (FCR) for birds received high PSP on wheat-based diets. Significantly heavier (P < 0.01) small intestine and gizzard + proventriculas weights at an early age, and heavier (P < 0.001) gizzard + proventriculas and pancreas during the grower stage were recorded in birds fed the corn-based diets. At 24 d, pancreatic chymotrypsin amidase and lipase enzymes were significantly (P < 0.01) more active in chickens fed the wheat-based diet. The interaction of PSP level and the type of the grain was significant (P < 0.01) for pancreatic chymotrypsin amidase and lipase as well as jejunal maltase (P < 0.05). Both experimental factors had a significant influence on jejunum histomorphology at 24 d of age (P < 0.001). There was a significant interaction between level of PSP and the type of grain (P < 0.05) resulting in the highest villus surface areas being seen for birds fed the corn-based diet and receiving the medium level of PSP. The trial demonstrated that PSP supplement can be included at between 50 and 100 g/kg of starter diets, depending on the basal diet.