A way to improve the effectiveness in e-learning is to offer the personalized approach to the learner. Adaptive e-learning system needs to use different strategies and technologies to predict and recommend the most likely preferred options for further learning material. This can be achieved by recommending and adapting the appearance of hyperlinks or simply by recommending actions and recourses. This paper presents an idea for integration of such recommender system into existing web-based Java tutoring system in order to provide various adaptive programming courses.