The aim of this study was to check the accuracy of laboratory methods to predict the apparent protein digestibility (CPd, %) and digestible protein content (DP, g/kgDM) of dog foods, avoiding the use of experimental animals in digestion trials. Twenty-eight commercial dry extruded dog foods were tested by three different methodologies: an adaptation of the in vitro incubation method described by Hervera et al. (J Anim Physiol Anim Nutr 2007, 91: 205) for estimation of digestible energy of commercial dog foods, a modification of the pH drop methodology proposed by Hsu et al. (J Food Sci 1977, 42: 1269) for protein evaluation of human foods and the Near Infra Red Spectroscopy (NIRS) technology. All the methods assessed revealed very good, similar accuracy in the prediction of DP content either using the in vitro method (r = 0.99; RSD = 6.76; CV = 2.31%), the pH drop method (r = 0.99; RSD = 6.94; CV = 3.02%) or the NIRS (R2 = 0.96; SECV = 10.50) method, although the in vitro digestion method showed the highest accuracy approach of in vivo crude protein apparent digestibility: CPd in vitro (r = 0.81; RSD = 2.01; CV = 2.41%); CPd pH-drop (r = 0.78; RSD = 2.48; CV = 2.98%) and NIRS (r2cv = 0.53; SECV = 2.37).