With the development of science and technology, library management informationization has gradually replaced the traditional management mode how to enable readers to efficiently select more suitable books for complicated books. Personalized recommendation algorithms have become particularly important. Our paper focuses on preprocessing borrowing data, analyzing the clustering algorithm and association rules, and exploring the potential laws that are hidden behind a large amount of data. In this way, a personalized reading recommendation model is constructed to provide readers with a more diverse and accurate recommendation model. In this paper, through clustering analysis and association rule analysis, it is found that different readers may have the same reading interests, and more readers will choose to borrow multiple books related to them at the same time when borrowing a certain book. Compared with the four algorithmic models ICF, qs-MAICF, jz-MAICF, and ms-MAICF, this paper recommends the most types of books for readers, which indicates that the model can tap the correlation between different books and provide readers with richer book choices. This paper is a significant accomplishment of artificial intelligence in the library industry, providing strong technical support for the modernization and intelligent development of libraries.