Despite a reduced estimation of MP supply (NRC, 2001) with CM inclusion, many studies reported a positive effect on milk production when canola meal (CM) substituted another protein source (PS), including soybean meal (SBM; JDS 96:1701). We hypothesized that mobile bag method underestimate CM intestinal digestibility (ID) of N estimated at 75% (NRC, 2001). Four cows equipped with ruminal and duodenal cannula were used in a randomized complete block design, each cow being a block. The PS evaluated were ¹⁵N-labeled CM (LCM), commercial CM (CCM) and commercial SBM. Each PS was rumen-incubated in nylon bags to determine DM and CP degradation and extra bags were incubated for 12 h to evaluate AA ruminal disappearance and produce 12-h residues. In each cow, 50 bags (10 and 20 μm pore size for CCM and SBM; and only 10 μm for LCM) of 12-h residues were inserted into the duodenal cannula and recovered in the feces to determine the ID. In addition, ID of LCM was estimated by direct introduction of 12-h residues into duodenal cannula; feces were totally collected for 28 h and ID = (1 - ¹⁵N recovered in feces/¹⁵N introduced into duodenal cannula). The effective degradability of CP, using a passage rate of 0.08 h⁻¹ and corrected for solubility, was similar among PS, averaging 63.5%, 62.8% and 63.1% (SEM = 2.5) for SBM, CCM and LCM, respectively. The ratio of essential AA (EAA) after 12-h of rumen incubation/raw EAA concentration (CP basis) was higher for CCM (1.22) than for SBM (1.02) and LCC (0.92) indicating the lowest degradability for CCM. The CP ID of LCM was in the same range with the mobile bag and the ¹⁵N methods, averaging 76.1% and 72.7% (SEM = 1.0), respectively. Contrariwise, there was a significant effect of pore size on the ID of CM: 64.7% with 10 μm vs 67.0% with 20 μm pore size, SEM = 0.1, indicating that a larger pore size allowed more material to leave the bag or more access for bacteria; the effect of pore size was not observed on SBM ID: 97.8% with 10 μm vs. 97.8% with 20 μm pore size, SEM = 0.7. Overall, this study showed that the reason for the underestimation of MP supply by NRC model (2001) in dairy ration including CM is likely not related to an underestimation of the ID of the RUP fraction of the CM.