The aim of this work was to characterize UV-absorbing compounds (UAC) in milk in relation to diet. In winter, 4 groups of cows each received a different diet: concentrate rich containing 35% cocksfoot hay (CCH), maize silage (MS), rye grass silage (RS), or rye grass hay (RH). In summer, 2 additional diets were given: mountain grassland hay (GH) and mountain grassland pasture (GP). Polyphenols were analyzed by HPLC and Folin reaction on forages and UAC were extracted from milks and analyzed by HPLC. In forages, the number of polyphenols was lowest in MS (57) and greatest in GP (85). Twenty-four peaks were identified, accounting for 28 to 47% peak area at 280 nm. Caffeoyl compounds and flavonoid glycosides were mainly found in RH, GH, and GP. Hydrolyzed compounds such as hydroxycinnamic acids and aglycones were found in MS and RS. Estimated amounts of polyphenols were lowest for MS (3.7 g/kg), roughly similar for CH, RS, and RH (about 15 g/kg), and greatest for GH and GP (21.6 and 35.3 g/kg, respectively). About 230 different peaks were separated in milks. Milks from RH and GP contained the lowest (87) and the greatest (127) numbers of peaks, respectively. Only 10 peaks were identified, accounting for 21 to 54% of the total spectra area. In addition to the major compound hippuric acid, phenylacetic acid, benzoic acid, 4-hydroxybenzoic acid benzaldehyde, catechol, and small amounts of ferulic acid were found in varying amounts depending on the diet. Flavonoids such as quercetin, luteolin, and apigenin were also present. Hippuric acid was clearly related to the presence of caffeoylquinic compounds in forages. Other identified UAC may originate essentially from forage simple polyphenols or from cell wall aromatics. Some of the several unknown compounds may also originate from the transformation of other nutrients. Estimated amounts of UAC were widely variable within each animal group. They were surprisingly high in CCH and roughly similar in all milks from preserved forages (about 3.6 mg/L), with generally greater values for GH milks, whereas the greatest amount was found in GP milks (13.3 mg/L). Hierarchical clustering clearly discriminated the 6 diets, showing that there were major differences in GP milks. Some UAC were specific to one or a group of diets. Ultraviolet-absorbing compounds are therefore a potential tool to distinguish between milks according to diet. In addition, they may have a bioactive effect on milk component conservation or on human health.