At present, artificial intelligence, block chaining, big data and other technologies set off the fourth industrial revolution. The persistence of COVID-19 has accelerated the promotion of digital and intelligent applications, and the big data has increased the difficulty of people's rapid access to demand, thus the recommendation system has been widely applied. The recommendation system has also been applied in the mainstream e-commerce platform in the book field. In this paper, through the collaborative filtering recommendation algorithm based on Mahout framework, combined with knowledge classification and user interest, a collaborative filtering recommendation algorithm based on knowledge classification and user interest is proposed. Through experimental analysis, the algorithm provides more refined personalized recommendation and higher recommendation quality.