The primary aim of this study was to assess the development of action planning in a group of typically developing children aged 3 to 10 years (N = 351). The second aim was to assess reliability of the action planning task and to relate the results of the action planning task to results of validated upper limb motor performance tests. Participants performed an action planning task in which they needed to grasp an object (a wooden play sword) and place it into a tight-fitting hole. Our main dependent variable was the grip type that participants used; that is, we measured whether initial grip was adapted in such a way that children reached a comfortable posture at the end of the action (the end-state comfort effect). Older children planned their actions more often in line with the end-state comfort effect compared with younger children. Test-retest and interrater reliability of the action planning task were good, with intraclass correlation coefficients (ICCs) of 0.90 and 0.95, respectively. We compared the action planning task with manual dexterity tests in a subset of participants (n = 197). We found a marginal relation with the manual dexterity tests, indicating that the action planning task measures different processes. In sum, our study showed that action planning increases from 3 to 10 years of age and that the experimental task we used is reliable in assessing anticipatory planning. Therefore, it may be used as a reliable additional test to investigate the degree to which motor behavior is affected at the cognitive level of anticipatory planning. (Contains 2 figures and 3 tables.)