With the rapid increase of digital resources, the time and cost for learners to filter out useful information from massive information also increase. By constructing a content-based and collaborative filtering digital resource recommendation model, we can improve the efficiency and accuracy of digital resource recommendation to learners. Then, from the four dimensions of learning motivation, self-efficacy, goal setting and environmental factors, using analytic hierarchy process, this paper analyzes the importance of each index to the effect of autonomous learning, and recommends the model of digital resources.