Problem Data analysis is the process of processing data with the aim of finding useful information that can be used as the basis for the analysis process. The data used to analyze a data is using SDN 2 Sarajaya Learner data, the author has a problem that there is no clustering of learner data using K-means at SDN 2 Sarajaya from this problem the author will group the data with the DBI value besides that it will look for what type of parameter is the best from the learner data at SDN 2 Sarajaya, namely with learner data using K-Means algorithm clustering with optimize parameters.Clustering is an algorithm for grouping some data into certain data groups (clusters). Student data has 150 data with 45 attributes including No, Name, NIPD, Jk, NISN, Birthplace, Birthdate, NIK, Religion, Address, Rt, Rw, Hamlet, Village, Subdistrict, Postal Code, Type of Residence, Transportation Equipment, KPS Acceptance, KPS No, Father's Name, Father's Year of Birth, Father's Last Education, Father's Occupation, Father's Income,Mother's Name,Mother's Birth Year,Mother's Last Education,Mother's Occupation,Mother's Income,Rombel,KIP Receipt,KIP Number,Birth Certificate Registration Number,Bank,Bank Account Number,Account in Name,KIP Eligibility (school proposal),Reason for KIP Eligibility,Special Needs,School of Origin,How Many Children,Latitude,Longitude. The tools used by researchers are rapidminer version 9.10.000. The K-Means algorithm is one of the algorithms used for partitioning because K-Means is based on determining the initial number of clusters by definition to find the initial value of the cetroid. The purpose of the research conducted by the author is to find information from the data. The test method used is the KnowlledgeDiscovry in Database (KDD) method. In the clustering results obtained by the learner data cluster using the Devies Bouldin Index value closest to 0 with cluster 2 to cluster 10 experiments resulting in the best K value in cluster 4 den DBI value 0.12 and the number of members L0 : 79 clusters L1: 1 cluster L2: 1 cluster L3: 69 clusters with parameter types BrigmanDivergences.