In order to offer methods for assessing the protein quality of dog foods, eight commercial dry-type complete dog foods were analysed by in vitro chemical methods (crude protein, amino acid composition, in vitro pepsin digestibility) and rat growth trial (weight gain = WG, feed efficiency = FE, protein efficiency ratio = PER, net protein ratio = NPR, and net protein utilisation = NPU). Three individual samples were taken from each dog food and their homogenous mixture was used as an ingredient for the test diets fed to six rats per diet. All dog food samples were used as a single protein source in isonitrogenous (10% crude protein) and isoenergetic (ME = 14.9 MJ/kg) complete diets. Rat weight gains showed significant differences among most of the test diets as a consequence of the differences in protein quality of the dog food samples. The protein quality indices of the dog food products showed significant variations. The measured ranges of protein quality indices of the samples were as follows: WG = -5.4-21.4 g; FE = 0-0.25 g/g; PER = 0-2.54 g/g; NPR = 1.06-3.52 g/g; NPU = 19.93-65.93%. The limiting amino acid, the sum of essential amino acids and the chemical scores were calculated from gross amino acid composition. Tryptophan was the first limiting amino acids in all samples but one. Methionine was found to be the second limiting amino acid. The crude protein content of the samples showed no correlation with indices of protein quality, indicating that the declaration of crude protein is not a sufficient information to judge the protein quality of dog foods. The declaration of NPR, or NPU indices may be recommended for labelling the protein quality of premium-type dog foods.