The difficulty often comes every year, in choosing graduation as a student by an employeeperengkrutan partners with AMIK Labuhanbatu. Competence of each student graduation takinginfluence the employee in question, so as not to disappoint both sides. In this study, the authors try toclassify the data using a competency-based graduation one simple data mining techniques, namelyclustering technique (clustering) two-dimensional, which means the two variables that will be used ingrouping the graduation GPA and the value of student competencies that. Algorithms used in thegrouping (clustering) using the K-Means algorithm that starts with a random selection of K, which isthe number of clusters to be formed from the data to be in the cluster. This testing is done manuallyin addition also performed with the RapidMiner data mining application of 46 records. From theresults of a study of 46 records were done manually or using RapidMiner data mining applicationshave 3 groups of competency-based graduation with similar results between the two tests, namelyCluster 1 consists of passing information to the student with IPK of 30.0 to 31.7 and 62.50 toCompetency with 71,50 the number of members of cluster 13. Cluster 2 consists of graduation studentswith IPK of 30.0 to 33.5 and 81.00 to 89.00 competencies that number 18 cluster members. Cluster 3consists of graduate students with IPK of 30.0 to 33.8 and 72.50 to 79.00 competencies that number15 cluster members.