Abstract: Adaptive Learning System (ALS) is an effective solution for the supporting service to personalized learning. Recommending personalized contents according to the differences of learners is the core mechanism of ALS. Based on the modeling of learners and domain knowledge, this study uses the content and collaborative filtering-based hybrid recommendation to realize a new-type personalized proactive recommendation mechanism. Experiment results show that the findings of this research could well recommend proper learning contents suitable for learners to improve their learning effect.