Background Most studies assessing the association between caesarean delivery (CD) and childhood overweight/obesity have failed to account for important confounders, such as maternal prepregnancy body mass index (BMI) or the indication of the CD. Furthermore, within-family analyses have reported contradictory results. We aimed at evaluating the association between CD and offspring's risk of overweight/obesity while adjusting for important confounders and accounting for correlations between siblings. Methods Women in the 'Seguimiento Universidad de Navarra' cohort provided structured information regarding their pregnancy history and their children's health through online cross-sectional questionnaires. We calculated adjusted differences in BMI z-score and risk ratios (RR) for offspring's overweight/obesity associated with CD, with hierarchical models to account for correlations between siblings. We also performed a within-family analysis in 341 siblings who were discordant in delivery mode, using conditional multivariable logistic regression. Results Among the 2791 children analysed, those born by CD had higher average BMI z-scores (difference: 0.17; 95% CI 0.07 to 0.27) and higher risk of overweight/obesity (RR: 1.32, 95% CI 1.05 to 1.65) than children born vaginally. The association did not differ by maternal characteristics or offspring's age strata, and the results were consistent in sensitivity analyses. Furthermore, within-family analysis showed that children born by CD had 2.67-fold higher risk of overweight/obesity (95% CI 1.10 to 5.12) than their peers born vaginally. Conclusion Children born by CD have higher average BMI z-scores and higher risk of overweight/obesity than children born vaginally. The consistency of these findings across multiple approaches to address potential residual confounding likely suggests a true biological effect.