With the rapid development of wireless communication and sensor technologies, ubiquitous learning has become a promising solution to educational problems. In context-aware ubiquitous learning environments, it is required that learning content is retrieved according to environmental contexts, such as learners' location. Also, a learning content retrieval scheme should be able to work with various instructional strategies for different learning activities. To solve the context-aware learning content retrieval problem, we propose a strategy-driven approach to derive content retrieval strategies from instructional strategies. Moreover, we construct a knowledge-based system to expand query keywords based on the derived strategies and then select relevant keywords according to geographical distance between entities of concept and learners. This system is composed of four components: knowledge transformation, query expansion, content retrieval and user interface. Besides, ontology construction algorithms, designed for teachers to easily build up ontology from course outlines, are applied to generate the rules of query expansion and the taxonomic index of learning object repository. The experimental results indicate that the proposed approach can increase the learning performance of students. (Contains 7 tables and 10 figures.)