Many modern pesticides have endocrine disrupting abilities and early-life exposure may affect growth and disease risk later in life. Previously, we reported associations between prenatal pesticide exposure and higher childhood body fat content measured by anthropometry. The associations were affected by child PON1 Q192R genotype. We aimed to study whether prenatal pesticide exposure was still associated with body fat content and distribution in the children at puberty and the potential impact of both maternal and child PON1 Q192R genotype. In this prospective cohort study of 247 children born by occupationally exposed or unexposed women (greenhouse workers and controls) two follow-up examinations (age 10–15 and 11–16 years) including simple anthropometry, skinfold measurements, pubertal staging and blood sampling were performed. Total and regional fat% was determined by dual X-ray absorptiometry (DXA) at age 10–15. Prenatal pesticide exposure was associated with increased total, android, and gynoid fat percentage (DXA) at age 10–15 years after adjustment for sex, socioeconomic status, and puberty (all β = 0.5 standard deviation score (SDS) p < 0.05). Stratified by sex, the associations were significant in girls (total fat: β = 0.7 SDS, android–gynoid ratio: β = 0.1, both p < 0.05), but not in boys. Carrying the R-allele (child or mother, separately, or both) augmented the differences between exposed and unexposed children (total fat: β = 1.0 SDS, β = 0.8 SDS, p < 0.05, respectively, and β = 1.2 SDS, p < 0.01). No exposure-related differences were found if either the child or mother had the QQ wild-type. At age 11–16, exposed children tended to have a higher total fat% estimated by skinfolds than unexposed children (p = 0.06). No significant associations between prenatal exposure and body mass index or waist circumference were found. Prenatal pesticide exposure was associated with higher adolescent body fat content, including android fat deposition, independent of puberty. Girls appeared more susceptible than boys. Furthermore, the association depended on maternal and child PON1 Q192R genotype.