BACKGROUND Childhood socioeconomic position (SEP) has previously been associated with increased risk of overweight among children and adolescents. However, it remains uncertain whether the timing of exposure is important in relation to developing overweight in early adulthood. We aimed to examine how SEP during early (0-8 years) and late childhood (9-14 years) relates to overweight at age 15, 18 and 21. METHODS Longitudinal study in Western Denmark of 2879 young people (aged 15 in 2004). Exposure variables from registers were yearly household income, parental highest educational level and parental labour market participation (LMP), supplemented with questionnaire information about "family functioning" (age 15). Outcome variables were overweight and obesity, measured at three-time points.We analyzed the adjusted associations between childhood SEP and overweight and obesity using multinomial logistic regression, stratified on gender. RESULTS Early childhood: Parental lower educational level increased girls' risk of overweight and obesity at age 18 and 21 between RR = 1.8 (95% CI 1.0;3.4) and RR = 5.2 (95% CI 1.4;19.3). Girls reporting poor "family functioning" had up to twice the risk of overweight and obesity at age 21. Boys, whose fathers had a lower level of education had up to 2.4 times the risk of obesity at age 21. Parental low LMP increased boys' risk of obesity at age 18 and 21 between RR = 2.2 (95% CI 1.3;3.8) and RR = 2.8 (95% CI 1.3;6.1). Late childhood: Parental lower level of education tripled the risk of overweight and obesity among girls at age 18 and among both genders at age 21. CONCLUSIONS This study confirmed to some extent that economic, social and psychological insecurity and inequality as measured by lower parental educational level, lower household income, low labour market participation and poor family function during childhood was associated with an increased risk of overweight and especially obesity in adolescence and early adulthood in both genders. Despite some imprecise measures, the direction of the associations pointed to several associations, which all were in the hypothesized direction. Timing of lower household income and parental low LMP in childhood seemed to be gender-specific in some way, but this warrants more studies.