Objective The current study prospectively examines the intra-uterine hypothesis by comparing maternal, paternal and grandparental lineage influences on children's diet and also maternal-child aggregation patterns during pregnancy and early childhood. Design Prenatal dietary information was available for expectant mothers, fathers and up to four grandparents through a detailed validated semi-quantitative FFQ. At 6-year follow-up, when children averaged 5 years of age, dietary information was re-collected for mothers and a subset of maternal grandmothers using the same FFQ. Child's FFQ version was used for children. Anthropometric and sociodemographic variables were also collected. Settings Three-generation familial cohort representative of the contemporary Irish national population. Subjects Children aged 5 years (n 567) and their parents and grandparents. Results Associations for energy, macronutrient and fibre intakes were compared using Pearson's correlations, intra-class correlations (ICC) and linear regression models, adjusted for energy and potential confounders. Significant, moderate-strength positive correlations were observed for nutrient intakes in children's nuclear families (ICC (range) = 0·22-0·28). The father-child associations (r (range) = 0·13-0·20) were weaker than the mother-child associations (r (range) = 0·14-0·33). In general, associations were stronger for maternal postnatal intake-child intake than for maternal prenatal intake-child intake, except for percentage of energy from fat (adjusted β = 0·16, 95 % CI 0·05, 0·26; P = 0·004), which was stronger for maternal prenatal intake, specifically in non-breast-fed children (adjusted β = 0·28, 95 % CI 0·12, 0·44; P = 0·001). Among all grandparents, correlations were significant only for maternal grandmother-mother pairs (r (range) = 0·10-0·36). Significant positive ICC were observed for nutrient intakes of maternal grandmother-mother- child triads (ICC (range) = 0·12-0·27), not found in paternal lines. Conclusions These findings suggest that maternal-environment programming influences dietary intake. Copyright © 2012 The Authors.