Background: Like adults, children need to allocate study time and endeavour optimally in order to enhance learning effectiveness. Aims: The aim of this study was to investigate the development of shifting from habitual to agenda-based processes on study decisions. Sample: The participants were 309 children in the second, fourth, and sixth grades. Methods: We adopted the research paradigm proposed by Ariel and Dunlosky (2013, "Mem. Cognit.," 41, 416). In Experiment 1, the students selected items to study either with or without time constraint. In Experiment 2, the students were instructed to select all three items or one item to study per trial. Results: The results of Experiment 1 showed that for sixth graders, the likelihood of selecting high-value items under the restricted-study-time condition was higher than that under the restricted-total-time condition; second and fourth graders failed to construct an agenda of prioritizing high-reward items for study in the restricted-study-time condition. In Experiment 2, when students were instructed to select one item to study per trial, high-value items were prioritized over items on the left of the array for sixth graders; on the contrary, second and fourth graders seemed to choose item for study arbitrary, although they tend to choose high-value items compared with their choice pattern in no-choice-limit condition. Conclusions: The results indicate that children can shift towards agenda-based process when habitual responding cannot maximize reward, and there is an age-related improvement in shifting between grade 4 and grade 6.