The purpose of this article is to explore the behaviors of readers by the way of data mining in order to attain new readers and promote the loyalty of existing readers as new model management in a library. Such kind of digital library really offers efficient and convenienct in terms of Customer Relation Management (CRM). Furthermore, there are many important issues on the application of readers’profiles such as personalized rec-ommendations or predictions based on readers’transactions. All of those recommenda-tions can be attained by data mining techniques. To some extent, data mining will be an embedded function for promoting high efficiency and effectiveness in the future digital library. This paper discusses the satisfaction of readers by CRM and the profiles of books and readers’borrowing behaviors by data mining techniques. Thus, the CRM can pro-vide readers highly efficient management and services. Data mining for readers at library will be both applied to collect and store digital documents, and provide a new paradigm shift as e-Learning web in the days to come.