The major cause of death in colorectal cancer is related to liver metastasis. Although the metastatic process has been well studied, many aspects of the molecular genetic basis of metastasis remain unclear. Elucidation of the molecular nature of liver metastasis is urgent to improve the outcome of colorectal cancer. We analyzed the chronological gene expression profiles of 104 colorectal samples corresponding to oncogenic development including normal mucosa, localized and metastasized primary tumors, and liver metastatic lesions as fundamental samples using a custom cDNA microarray. The gene expression patterns in 104 samples were classified into four groups closely associated with their metastatic status, and the genes of each group appropriately reflected the metastatic process. To investigate the existence of metastatic potential in primary tumors using metastasis-related genes detected by chronological analysis, we performed a hierarchical cluster and supervised classification analysis of 28 independent primary tumors. Hierarchical cluster analysis segregated the tumors according to their final metastatic status, rather than their clinical stages, and the profile of metastasized primary tumors resembled one of a metastatic lesion apart from a primary lesion rather than one of a non-metastasized primary tumor. Using the supervised classification approach, the expression profile of these genes allowed the classification of tumors diagnosed as localized cancer into two classes, the localized and the metastasized class, according to their final metastatic status. The disease-free survival and overall survival were significantly longer in the localized class than the metastasized class. Chronological analysis of the gene expression profile provides a better understanding of the metastatic process. Our results suggest that the metastatic potential is already encoded in the primary tumor and is detectable by a gene expression profile, which allows the prediction of liver metastasis in patients diagnosed with localized tumors and also the design of new strategies for the treatment and diagnosis of colorectal cancer.