<p>Urban biodiversity loss and rapid urbanization pose significant challenges to architectural education, highlighting a gap in integrating ecological knowledge into design practices. This study develops and validates the Knowledge graph for pattern language in urban biodiversity, an educational framework that combines pattern language theory, knowledge graphs, and large language models to enhance ecological integration in architectural teaching. A five-phase mixed-methods approach is employed: first, a systematic literature review; second, the development of design patterns inspired by Christopher Alexander; third, the creation of an ontology and the construction of a knowledge graph, integrated with a language model to generate design recommendations tailored to urban biodiversity contexts; fourth, validation through participatory workshops with students; and fifth, expert evaluations that confirmed the framework’s effectiveness in incorporating biodiversity criteria into architectural design. Although results are currently limited to academic environments, the framework offers a scalable approach to AI-assisted architectural education.</p>