BACKGROUND & AIMS: Programming of infant development and later health may depend on early-milk polyunsaturated fatty acids (PUFA) contents, that are very variable between women for reasons not well elucidated. Indeed, a high n-6/n-3 PUFA in milk was associated with higher adiposity, arterial pressure and lower psychomotor scores in childhood. We aimed to explore the respective contribution of several maternal and perinatal factors to the variability of linoleic (LA), α-linolenic (ALA), arachidonic (AA), and docosahexaenoic (DHA) acid levels in early milk. METHODS: Fatty acids of 934 colostrum samples from the EDEN mother-child cohort were analyzed by gas chromatography. The dietary intakes during the last trimester of pregnancy were estimated using a quantitative food frequency questionnaire. Relationship between milk PUFA and dietary fatty acids, and other maternal or pregnancy variables were analyzed by multiple linear regression. RESULTS: The means (±SD) of colostrum LA, ALA, AA and DHA levels were, respectively, 9.85 ± 1.85, 0.65 ± 0.22, 0.86 ± 0.16, and 0.64 ± 0.19% of total fatty acids. Obese mothers colostrum contained the highest level of LA and AA and the lowest level of ALA and DHA. Colostrum LA, AA and DHA levels were higher in primiparous women. Mother's age was positively associated with colostrum AA and DHA. Dietary n-6 PUFA were associated with higher LA and lower DHA levels in colostrum, while dietary n-3 PUFA were related to higher LA and lower AA levels. Contrary to what was observed for DHA, AA level in colostrum was not related to its dietary intake. High dietary AA/DHA and total n-6/n-3 ratios were critical for the content of DHA in colostrum lipids. CONCLUSIONS: Our study brings new insights in the understanding of the main maternal factors involved in PUFA levels variability in early milk. These data are important to consider for dietary counseling for women prior to and during pregnancy.