The process of compiling books on bookshelves at the University of Muhammadiyah Kotabumi has not been based on borrowing data. The pattern used so far is still in grouping books based on the availability of books. This sometimes makes it difficult for members who want to borrow books. The borrower must first browse through the shelves quite a lot to find the book to be borrowed. To overcome the existing problems, it is necessary to carry out book data management techniques based on the tendency of members to borrow several book items on adjacent shelves. One technique that can be used by library managers is the data mining technique with the FP-Growth algorithm. This research is intended to assist library managers in processing loan data and determining patterns for arranging the layout of books on shelves based on borrowed data. Implementation of data mining using the FP-Growth rule algorithm in the test assisted by Rapidminer tools.