Grounded in the principles of developmental psychology, this study develops an innovative online educational game for young children titled “Fill in the Picture.” Utilizing the AppCan platform in conjunction with Photoshop technology, the game involves the automatic generation of randomly sampled, unordered point sets using network depth search algorithms. It captures the sequential order in which children connect these points onscreen to form coherent graphical representations, subsequently assessing the correctness of these connections against computer-judged standards. The effectiveness of “Fill in the Picture” was evaluated through a questionnaire survey, which revealed a weighting coefficient of 0.248 for the game’s educational quality. Additionally, it scored above 4 on various dimensions, including educational value, scientific basis, playfulness, artistic integration, immersion, and technological engagement. Importantly, in terms of pedagogical design satisfaction, 81.3% of respondents recognized that the game significantly enhances understanding and mastery of scientific concepts, while 81% appreciated the logical organization of its content. Thus, the “Fill in the Picture” game, designed and developed by integrating a tailored educational game design model based on developmental psychology, demonstrates substantial educational efficacy and user satisfaction.