Uveal (ocular) melanoma is a highly aggressive cancer that leads to metastatic death in up to half of patients despite successful local therapy. Biomarkers of metastatic risk are critically needed to institute new adjuvant treatment strategies in high-risk patients. Previously, we showed that two prognostically significant molecular subtypes of uveal melanoma could be identified based on gene expression profiling of the primary tumor. In this study, we investigated the value of micro-RNA (miRNA) expression patterns in predicting metastatic risk. A genome-wide, microarray-based approach was used to screen for differentially expressed miRNAs using the Agilent miRNA microarray (Agilent Technologies, Foster City, California, USA) platform containing probes for 470 human miRNAs. Unsupervised analysis was performed using principal component analysis, and supervised analysis was performed using significance analysis of microarrays. Tumors readily clustered based on miRNA expression into two groups that corresponded to the gene expression-based subtypes: class 1 (low metastatic risk) and class 2 (high metastatic risk). The most significant discriminators were let-7b and miR-199a, and the expression of these miRNAs was validated by quantitative PCR. A classifier that included the top six miRNA discriminators accurately distinguished class 1 from class 2 tumors with 100% sensitivity and specificity. miRNA expression may represent a highly accurate biomarker for metastatic risk in uveal melanoma. In addition, these results may provide new insights into the role of miRNAs in tumor progression and the metastatic phenotype.