Background Childhood overweight/obesity has been associated with environmental context, such as green space, gardens, crime and deprivation. This paper assesses the longitudinal association between environment and body mass index (BMI) for children across the ages of 3-11 years. It also investigates the relationship between environment and child overweight/obesity Methods 6001 children from the UK Millennium Cohort Study living in England were analysed. We estimated fixed effects linear and logistic regression models of the association between environment (levels of green space, gardens, crime and deprivation) and BMI/overweight of children at four time points between the ages of 3 and 11. Models were adjusted for age-related changes in weight, child sex and education level of the main carer. Results Statistically significant associations were found between environmental measures of both more gardens and lower levels of crime and lower BMI (effect size (95% CI) respectively: 0.02 (0.04 to 0.00), 0.04 (0.07 to 0.02)). Areas with less crime were associated with a slightly lower odds of overweight among children with a higher educated parent (OR 0.93 (0.87-0.99)) Conclusions By exploiting longitudinal measures of environment and BMI this study is able to establish a more causal association between environment and BMI. Environments with more gardens and lower crime tend to result in slightly lower BMI. However, the effect sizes are small and non-significant odds of changing weight status do not support environmental factors as a key determinant of cohort changes in childhood overweight/obesity.