The objective of this paper is to analyze the relationship among GIS terms using a text mining technique. The procedure was: first, every page of twenty one GIS textbooks was scanned and digital texts were obtained; second, the 105 most frequently used terms were selected from the index terms in the twenty one textbooks; last, the105 terms were clustered according to co-occurrence information in the body texts of the twenty one textbooks. As a result, classes of the terms were obtained by objective analyses.