Background: Parental restrictive feeding (i.e., limiting food intake of children) has been linked to childhood overweight. However, the directionality of the causal pathway remains unknown. Objective: The objectives of this study were to examine the bidirectional association of maternal restrictive feeding with children's weight and body composition across childhood and to explore a possible mediating role of maternal concern about child weight. Design: Data were available for 4689 mother-child dyads participating in Generation R, a prospective birth cohort in the Netherlands. At ages 4 and 10 y, restrictive feeding was assessed with the parent-reported Child Feeding Questionnaire, and children's body mass index (BMI) was measured. At age 6 y, fat mass index (FMI) and fat-free mass index (FFMI) were measured with dual-energy X-ray absorptiometry. Both directions of the relation between restriction and child body composition were examined with multivariable linear regression analyses and cross-lagged modeling. Mediation analyses were performed to examine concern about child weight (mother reported at child age of 10 y) as a potential mediator. Results: Higher child sex-and age-adjusted BMI SD scores (zBMI) at age 4 y predicted more restrictive feeding at age 10 y (B = 0.15; 95% CI: 0.11, 0.18). Both sex-and age-adjusted FMI SD scores (zFMI) and sex-and age-adjusted FFMI SD scores (zFFMI) at 6 y were also positively associated with restrictive feeding at 10 y. Maternal concern about child weight partially mediated these associations from child body composition to restrictive feeding (e.g., for zBMI at 4 y: B indirect = 0.10; 95% CI: 0.07, 0.13). There was no temporal association from restrictive feeding at age 4 y to child zBMI at age 10 y after adjustment for baseline zBMI. Conclusions: The continued use of restrictive feeding practices at age 10 y appeared to be primarily a response of mothers to an unhealthy weight of their child rather than a cause of children's overweight. Guidelines discouraging restrictive feeding for preventing childhood overweight should therefore be reconsidered. Am J Clin Nutr 2017;106:783-90.