The growing demand for precise and scientific exercise knowledge and personalized training guidance has highlighted the urgent need for standardized organization and intelligent expression of sports and exercise knowledge. This study addresses these needs by utilizing ontology theory to collect and mine multi-source, complex sports and exercise knowledge at a large scale. We establish a hierarchical classification system for sports and exercise knowledge and develop an ontology-based knowledge representation model to structure the data semantically. By integrating and deeply mining fragmented and cross-disciplinary sports and exercise knowledge, we construct a sports and exercise knowledge corpus that facilitates centralized integration, shared analysis, and full utilization of sports and exercise knowledge data. Furthermore, using this knowledge corpus, we explore key technologies for an intelligent query system, enabling intelligent management, querying, and Q&A functionalities for sports and exercise knowledge. This knowledge graph can meet the public’s diverse and personalized needs for sports and exercise knowledge, thereby supporting targeted and effective exercise and workouts.