Background: In highly developed countries, childhood overweight and many overweight-related risk factors are negatively associated with socioeconomic status (SES). Objective: The objective of this study is to investigate the longitudinal association between parental SES and childhood overweight, and to clarify whether familial, psychosocial or behavioural factors can explain any SES gradient. Methods: The baseline and follow-up surveys of the identification and prevention of dietary and lifestyle induced health effects in children and infants (IDEFICS) study are used to investigate the longitudinal association between SES, familial, psychosocial and behavioural factors, and the prevalence of childhood overweight. A total of 5819 children (50.5% boys and 49.5% girls) were included. Results: The risk for being overweight after 2 years at follow-up in children who were non-overweight at baseline increases with a lower SES. For children who were initially overweight, a lower parental SES carries a lower probability for a non-overweight weight status at follow-up. The effect of parental SES is only moderately attenuated by single familial, psychosocial or behavioural factors; however, it can be fully explained by their combined effect. Most influential of the investigated risk factors were feeding/eating practices, parental body mass index, physical activity behaviour and proportion of sedentary activity. Conclusion: Prevention strategies for childhood overweight should focus on actual behaviours, whereas acknowledging that these behaviours are more prevalent in lower SES families.