The OPALINE project aimed at understanding the determining factors of the development of food preferences and eating behaviour up to the age of 2 years by following a cohort of children with a longitudinal recording of perinatal and postnatal feeding experiences, of children’s sensitivity to food tastes and odours and of parental feeding practices. The aim was to conjointly analyse the datasets to draw an overall picture of these potential determinants of food preferences and of their relative weight over the course of the first two years. The recruitment of a cohort of children (N=314) was conducted thanks to the help of local health and childhood professionals. The milk, complementary and table feeding received by each child and by his/her mother during pregnancy and lactation was recorded using food diaries. The acceptance of food tastes and odours was evaluated at different ages in the laboratory. The family environment was evaluated through questionnaires (parental feeding style; strategies when the child refuses to taste a food; attitudes or criteria to buy foods for children). Food preferences regarding the past three months were measured using questionnaires at 15, 18, 21 and 24 months old. Analyses were based on the Structural Equation Modelling method, to model the causal relationships between these factors and food preferences at weaning and up to the age of 2 years. This presentation will mainly focus on the determinants of liking of vegetables (LVeg), the least liked food category at 2 years. Liking of vegetables at weaning was positively correlated with the variety of vegetables introduced at the beginning of weaning, but it was correlated neither with breastfeeding duration nor with variety of vegetables consumed by the mother during pregnancy (or lactation), and nor with age at weaning. At one year, LVeg was positively correlated with acceptance of bitter taste and of two vegetable odours, 2-isobutyl-3-methoxypyrazine and dimethyl disulphide. LVeg at weaning was positively correlated with LVeg at the later ages up to 2 years. LVeg during the second year was strongly related with the frequency of exposure to vegetables, but surprisingly negatively related with variety. Finally, at 2 years, “permissive” parental feeding strategies were associated to a lower LVeg. This modelling approach makes it possible to evaluate the respective weight of individual factors (acceptance of some flavours), and of parental feeding practices and strategies, i.e. the way to conduct weaning, and their strategies when the child refuses to taste a food in determining children’s food preferences.