With the development of network technology, university library construction is gradually moving in the direction of digital libraries; digital age of university libraries has arrived. By automated information management system in libraries, readers can easily inquire book information and borrow books, as well as manage bibliographic information retrieval and WEB access record query. In order to know how to analyze and predict readersÂ needs and saving trends, data mining technology is introduced, through which readers can easily find the useful information which can be utilized from mass of data. This paper begins with the basic concept of data mining, systematically introduces its concept and implementation methods, and then analyzes the related knowledge of personalized information system in university libraries accompanied with comparative analysis of its development status. It is found that university libraries lack the ability of data mining. Based on the circumstance, this study designed a set of model structure combined with library system in order to achieve personalized information system based on data mining technology. Then the study analyzed the composition of each module from a theoretical point of view as well as the explanation of functional description and workflow. Model building provides a better theoretical basis for the development of university libraries, thereby improves application ability of university libraries.