Human infants are preferably fed human milk (HM), but a majority still receive infant formula (IF) as a HM substitute. Optimization of IF is still required to improve HM biomimetics, including digestion behavior biomimetics. While this can be studied in vivo, such experiments have to be reduced and appropriate in vitro models are needed. The present study aimed to compare food deconstruction and protein digestion of HM vs. IF using two infant digestion models, the mini-piglet and the in vitro dynamic gastrointestinal system (DIDGI®). Mini-piglets (Yucatan) were fed either a mature HM (n=9) or a standard IF (n=9) during 6 days. Piglet digesta were collected along the digestive tract 30 min after the last meal. The same foods were digested in triplicate using a term infant in vitro dynamic model with regular digesta sampling along time. Microstructure (confocal microscopy and laser diffraction) and protein digestion (SDS-PAGE, hydrolysis degree, peptidomics) were investigated in both digestion models. Data were statistically analyzed thanks to ANOVA and multidimensional analyses (hierarchical classification and multiple factor analyses). The microstructure of the digesta differed between HM and IF in a similar manner in vitro and in vivo along digestion. The meal dilution and emptying were similar between both digestion models, with a faster emptying for HM. Proteolysis, as investigated by SDS-PAGE, were similar between digestion models, with a lower hydrolysis level for HM caseins. Peptide mapping along the sequence of the major proteins was well correlated between models, particularly in the stomach and the proximal jejunum (r > 0.6). Similar result was found for bioactive peptide release. The ratio between bioaccessibility (in vitro) and bioavailability (in vivo) of amino acids was high (50-80%) at the cleavage sites of the pancreatic enzymes, more precisely for Arg, Tyr, Lys, Phe and Leu, but was much lower for the other amino acids (<30%). In overall, the in vitro dynamic gastrointestinal digestion model well predicted the in vivo digestion of HM and IF, particularly for protein hydrolysis, peptidomics and food deconstruction, while further improvement is needed to better correlate bioaccessibility and bioavailability.