The use of biomarkers of food intake in serum or urine has shown great promise for completing traditional dietary assessment tools. Untargeted metabolomics was used here to identify new biomarkers of milk and cheese intake and further explore the metabolic response to the ingestion of such foods. A randomized controlled crossover study was conducted in eleven healthy men and women. Serum and urine samples were taken postprandially up to 6 h and fasting 24 h after a single intake of isocaloric doses of milk, cheese or soy drink (non-dairy control). Untargeted metabolomics was conducted using UHPLC-MS. Biomarkers of intake were first selected in serum by multivariate statistical analysis, their mass distribution and postprandial kinetics were characterized, and their presence in urine was then investigated. Among the 2988 unique metabolites detected in serum, 261 were selected as potential biomarkers of milk, cheese or soy intake. Biomarkers of cheese intake had a significant lower mass distribution than biomarkers of milk or soy intake. The identities of nine compounds were confirmed (5 biomarkers of cheese intake and 4 of milk intake) including saccharides, amino acids, amino acids derivatives and dipeptides. Among them, six were also found in urine. Two oligosaccharides, the blood group H disaccharide and the Lewis a trisaccharide, appeared as clear serum biomarkers of milk intake but with high inter-individual variability. Interestingly, the two oligosaccharides presented opposite trends, i.e. after milk intake, subjects showing an increase in blood group H disaccharide did not show any increase in Lewis a trisaccharide, and conversely. This result was confirmed in urine. New biomarkers of milk or cheese intake could be identified using HPLC-MS based untargeted metabolomics, most of them being related to peptides and amino-acids metabolism. The observations made for the blood group H disaccharide and the Lewis a trisaccharide first recalls the importance of inter-individual variability when searching for biomarkers of intake, but also encourages further investigations towards the role of the Lewis antigen system and related genotypes in the differential response to milk intake and the potential health outcomes.