Ovarian cancer represents one of the most malignant gynecological tumors. Despite recent advances in treatment, ovarian cancer remains to be highly susceptible to metastasis. However, information concerning genome-wide gene expression profiles is limited to develop a metastasis-specific gene signature in ovarian cancer. In this work, we try to identify changes in gene expression profile that underlie ovarian cancer metastasis. The dataset GSE73168 deposited in the Gene Expression Omnibus (GEO) database was processed to identify differentially expressed genes (DEGs) between primary tumor and metastatic tumor samples. The weighted gene correlation network analysis (WGCNA) was conducted for modules related to ovarian cancer metastasis. Modular genes associated with ovarian cancer metastasis were summarized for the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Receiver operating characteristic (ROC) curves were plotted to estimate the superiority of candidate genes in detecting ovarian cancer metastasis. The WGCNA yielded 25 co-expression network modules in the dataset GSE73168, and highly correlated genes with ovarian cancer metastasis were identified in the blue module. Twenty-two genes demonstrated differential expression between primary tumor and metastatic tumor samples, and two downregulated genes (P2RY13 and NKX6-1) and three upregulated genes (CD36, LOC57399 and RP11-587D21.4) of these 22 DEGs was also shown to correlate with ovarian cancer metastasis in the blue module. The area under the ROC curve verified these five DEGs as metastasis-specific genes for ovarian cancer. These results show P2RY13, NKX6-1, CD36, LOC57399 and RP11-587D21.4 serve as metastasis-specific genes for ovarian cancer.