Cowpea might have been introduced from China to Korea and cultivated for several hundred years but it has never been a staple food crop in Korea. In this study, genetic diversity of 492 Korean cowpea landrace accessions that have passport information was estimated using six SSR markers. The mean of Weir's gene diversity was 0.665 from all accessions investigated in the study. Cowpea gene diversity of six local provinces in Korea was ranged from 0.370 in accessions of Gangwon to 0.680 in Jeonra provinces. Low gene diversity of the cowpea genepool of Gangwon province was probably derived from relatively few introductions. Especially SSR markers VM36 and VM39 seem to be good markers to distinguish the Gangwon accessions from others by occurring at a specific locus with higher than 78% of allele frequency. Except for the Gangwon province with the low genetic diversity, gene diversity of cowpea accessions from other provinces was ranged from 0.600 to 0.680 indicating no big differences among provinces. Distribution pattern of the allele frequencies was similar among the other provinces. This may reveal that Korean farmers might exchange cowpea seeds easily with even their neighbors with geographical barriers. A core collection, 100 landraces, ca. 20% of base collection, was developed at the 70% of a similarity coefficient level using random sampling approaches after stratification of the entire landrace collection based on the phenetic dendrogram. The variability of SSR in the base and core collections of Korean cowpea landrace was compared by calculating Weir's gene diversity. The mean of Weir's gene diversity of the core was 0.707 while that of the base collection was 0.665. The higher diversity index in the core collection indicates that it maintains the initial variability and well represents the base collection. The core collection included one of determinate accession (IT 216155) and two of no branching type accessions (IT 103959 and IT 161024). The core collection could be used to guide more efficient management and utilization of the entire collection. This core collection should be revised periodically as additional accessions are collected and further characterization is conducted.