meet the growing demand in the field of book recommendation, the research focuses on meeting the personalized needs, behavioral patterns, and interests of readers. A book recommendation algorithm that combines K-means clustering with time information is proposed to provide more convenient and efficient book recommendation services and enhance readers' reading experience. The algorithm constructs a comprehensive user preference matrix by incorporating readers' borrowing time. Then, the K-means clustering is applied to group users with similar preferences and leverages a latent factor model to train and predict user ratings. The methodological integration of clustering and latent factor model ensures a more precise and dynamic recommendation process. The experimental results demonstrated that the proposed algorithm achieved a high average recommendation accuracy of 98.7%. Additionally, the algorithm maintained an average book popularity score of 8.2 after reaching stability, indicating its ability to suggest widely appreciated books. These outcomes validate the effectiveness of the algorithm in delivering accurate and popular book recommendations tailored to individual readers' needs. This study combines K-means clustering with time sensitive preference analysis and latent factor model to introduce an innovative method in the field of book recommendation systems. The findings provide valuable insights and practical applications for libraries seeking to enhance their personalized recommendation services, offering a significant contribution to the field of intelligent information retrieval.