Artificial Intelligence (AI) technology is reshaping the landscape of foreign language teaching, driving a paradigm shift from standardized, collective instruction to a learner-centered, data-driven model characterized by precision and personalization. However, current research and practice exhibit a significant imbalance across languages, with most advancements concentrated in English teaching, while studies on AI applications for less-commonly-taught languages like Japanese remain nascent. This study addresses this gap by constructing and applying an AI-enabled teaching model specifically for Japanese grammar instruction. The model is designed to address inherent challenges in traditional Japanese grammar teaching, such as the tension between collective instruction and individual learning needs, the disconnection between memorizing rules and applying them, and the lack of personalized, timely feedback in consolidation phases. Leveraging the core functionalities of the Zhihuishu platform (e.g., course setup, resource push, data statistics) and the natural language processing capabilities of AI tools like DeepSeek and Doubao, the model follows a three-stage progressive structure: “Pre-class Diagnosis–In-class Internalization–Post-class Consolidation”, forming a complete teaching loop supported by a data feedback mechanism. A detailed case study on the challenging “Honorifics System” demonstrates the model’s practical implementation, showing how data-driven diagnosis, interactive scenario-based exploration, and personalized consolidation tasks are operationalized. Practice indicates that this model effectively mitigates the shortcomings of traditional teaching in personalized tutoring, immediate feedback, and application transfer, significantly enhancing students’ grammar mastery and usage capabilities. The study concludes that the AI-enabled model provides a viable, operational solution for achieving precise, contextualized, and personalized Japanese grammar teaching, while also pointing to future directions for optimizing AI’s analytical capacity for complex grammar and pragmatic nuances and balancing technological integration with the teacher’s guiding role.