Maternal feeding is a frequent intervention target for the prevention of early childhood obesity but longitudinal associations between feeding and child overweight are poorly understood. This observational cohort study sought to examine the cross-lagged associations between maternal feeding and overweight across ages 21, 27, and 33 months. Feeding was measured by maternal self-report (n = 222) at each age. Child weight and length were measured. Cross-lagged analysis was used to evaluate longitudinal associations between feeding and overweight, adjusting for infant birth weight, maternal body mass index, maternal education, and maternal depressive symptoms. The sample was 50.5% white, 52.3% male and 37.8% of mothers had a high school education or less. A total of 30.6%, 29.2%, and 26.3% of the sample was overweight at each age, respectively. Pressuring to Finish, Restrictive with regard to Amount, Restrictive with regard to Diet Quality, Laissez-Faire with regard to Diet Quality, Responsiveness to Satiety, Indulgent Permissive, Indulgent Coaxing, Indulgent Soothing, and Indulgent Pampering each tracked strongly across toddlerhood. There were no significant associations between maternal feeding and child overweight either in cross-sectional or cross-lagged associations. Our results do not support a strong causal role for feeding in childhood overweight. Future work longitudinal work should consider alternative approaches to conceptualizing feeding and alternative measurement approaches.