This study investigated age‐dependent improvements of monitoring and control in 7/8‐ and 9/10‐year‐old children. We addressed prospective (judgments of learning and restudy selections) and retrospective metacognitive skills (confidence judgments and withdrawal of answers). Children (N = 305) completed a paired‐associate learning task twice, with a 1‐year delay. Results revealed improvements in retrospective, but not in prospective monitoring and control. Furthermore, control remained suboptimal, seemingly a consequence of overoptimistic monitoring. Both age groups showed stronger monitoring‐based control at the second compared to the first assessment. The comparison with a cross‐sectional sample (N = 144) revealed that improvements in retrospective monitoring can be mainly attributed to naturally occurring development, whereas retrospective control seemed to improve due to increased task familiarity.