This study addresses the early screening and personalized intervention of Functional Articulation Disorders (FAD) in children by innovatively applying artificial intelligence (AI) technologies to improve diagnostic accuracy and rehabilitation efficiency. The project, based at Shanghai Children’s Hospital, constructed a Mandarin children’s articulation disorder dataset and developed a rehabilitation-oriented diagnostic classification model. By integrating data-knowledge fusion, deep learning algorithms, and digital interactive rehabilitation games, the system achieves precise assessment and personalized rehabilitation training for children with articulation disorders. Experimental results show that the model achieves a diagnostic accuracy of 97.052%, significantly improving rehabilitation efficiency and patient compliance. The research outcomes hold significant medical, economic, and social value, providing scientific support for children’s speech health protection.