While previous studies have suggested that there exists a relationship between obesity and executive function (EF), the mechanisms and causal relationship between them remain unclear. There are important clinical implications of determining whether EF can predict and treat obesity. We conducted a multilevel meta-analysis of randomized controlled trials (RCTs) and longitudinal studies. Specifically, we investigate (a) whether EF interventions have an effect on weight loss, (b) whether baseline EF can be a predictor of future weight loss through obesity intervention, and (c) whether early-life EF can predict future weight loss. Eight RCTs and 17 longitudinal studies with a total of 11,393 participants were identified. We found that (a) EF interventions may not have an effect on weight loss, (b) baseline inhibition (β = 0.259, p = 0.03) and delay discounting (β = −0.17, p = 0.04) significantly predict future weight loss through obesity intervention, (c) age (F = 13.666, p = 0.005) moderates the relationship between working memory and weight loss through intervention, but not weight status, type of intervention, and percentage of female, and (d) early life inhibition (β = 0.185, p = 0.07) is a marginally significant predictor of future weight loss. Our results seem to support the assumption that the relationship between EF and obesity is not direct, and a higher-order factor, such as genes, may link obesity and EF. Building on the preliminary findings, further studies focusing on EF and obesity are needed in the future.