The development of new sequencing technology has stimulated the cancer-related genome analysis. Copy number variation is one of the most important features that represents the structural variation. In this paper, we suggest the metastasis identification method of melanoma using copy number variations. The identification marker is defined in consideration of the presence and the type of copy number variations in primary and metastatic tumors. The optimization of marker is also provided and classification performances of developed markers are compared using the linear classifier.