Breastfeeding has been associated with improved cognitive development. This may be explained by polyunsaturated fatty acid (PUFA) content of breast milk, especially long-chain (LC) PUFA that are needed for postnatal brain growth. Using data from the French EDEN cohort, we aimed to study whether the PUFA content of colostrum may explain observed associations between breastfeeding duration and cognitive scores at 2 and 3 y. A total of 709 breastfed children with available data on PUFA composition of milk were assessed using parent-reported questionnaires for motor and language at 2 y of age, or global cognition at 3 y. Multiple linear regressions were used to examine associations between PUFA levels and child cognitive scores, after controlling for many confounders. We found no association between LCPUFA levels in colostrum and child development. However, levels of linoleic acid (LA) were negatively associated with motor and cognitive scores, independently of breastfeeding duration. Children breastfed with the highest levels of LA tended to score closer to the never breastfed children than children breastfed with the lowest levels of LA. Our findings suggest that too high levels of LA in colostrum are associated with poorer child development at 2 and 3 y.