Human milk oligosaccharides (HMOs) are complex glycans composed of 5 different monosaccharides assembled in highly diverse structures of 3 to more than 20 sugar units, with many isomers. While more than 200 different structures of HMOs have been described in literature, an individual milk sample may contain 50 to 130 different HMOs. In order to describe the global composition of HMOs, their inter-individual variability, the impact of environmental factors on their relative concentrations, and their association with health outcomes in childhood, we analysed 317 early breast milk samples, collected 1 to 6 days after delivery in the Mother-Child EDEN cohort, using liquid chromatography coupled to high-resolution mass spectrometry (LC-HRMS, Q-TOF instrument, Bruker). We developed a fully automated R workflow to perform a semi-targeted analy- sis of all potential HMOs detected within the 317 raw LC-HRMS results: it covers the pre-processing of the spectra (with fine tuning of XCMS package parameters), and the annotation of HMOs based on both monosaccharide composition determination thanks to high-mass accuracy measurement and the availability of 19 commercial standards. From more than 20,000 initially detected peaks, 350 distinct HMO compositions (cor- responding to 634 potential HMOs including isomers) were retained after quality filter application. Among those, 18 were annotated by comparison with pure authentic stan- dards. These results led to the classification of the 317 mothers in four human milk groups related to Secretor and Lewis genotypes. It represents the first step in our reviewing process of HMO composition in early breast milk and its association with child health.