Internal models provide a coherent framework for understanding motor behavior. Examples for the use of internal models include anticipatory postural adjustments (APAs), where the individual anticipates and cancels out the destabilizing effect of movement on body posture. Yet little is known about the functional changes in the brain supporting the development of APAs. Here, we addressed this issue by relating individual differences in APAs as assessed during bimanual load lifting to interindividual variation in brain network interactions at rest. We showed that the strength of the connectivity between three main canonical brain networks, namely the cingulo-opercular, the fronto-parietal and the somatosensory-motor networks, is an index of the ability to implement APAs from late childhood (9- to 11-year-old children). We also found an effect of age on the relationship between APAs and coupling strength between these networks, consistent with the notion that APAs are near but not yet fully mature in children. We discuss the implications of these findings for our understanding of learning disorders with impairment in predictive motor control.