Children’s drawings serve as a vital window into their cognitive and emotional development, offering profound insights into their perspectives, creativity, and psychological states. However, traditional evaluation methods, reliant on subjective expert judgment, are often inconsistent, time-consuming, and impractical for large-scale educational applications. These limitations hinder the ability to provide timely, objective feedback, creating a critical gap in leveraging children’s artistic expressions for educational and developmental purposes. To address these challenges, we introduce DrawAA-Net, an AI-supported framework designed to automate the evaluation of children’s drawings. By integrating advanced deep learning and computer vision techniques, DrawAA-Net assesses multiple dimensions of artistic expression, including technical skills and aesthetic quality. Through extensive experimentation, we demonstrate that DrawAA-Net provides consistent, real-time feedback while balancing scientific rigor with human aesthetic preferences. This innovative tool not only enhances the efficiency and objectivity of drawing evaluations but also supports educators in fostering children’s cognitive and emotional growth. Furthermore, by automating the assessment process, DrawAA-Net empowers educators to unlock the full potential of children’s artistic expressions, offering a scalable solution for educational practice and developmental research. Project page: https://github.com/Nancywsn/DrawAA .