PURPOSE: MicroRNAs (miRNAs) are non-coding RNAs which have attracted attention as biomarkers in a variety of diseases. However, extensive unbiased analysis of miRNA in vitreous humor of sarcoidosis patients has not been reported. In the present study, we comprehensively analyzed the dysregulated miRNAs in ocular sarcoidosis to search for potential biomarkers. MATERIALS AND METHODS: This study included seven patients diagnosed with ocular sarcoidosis (five definite and two presumed). Five patients with unclassified uveitis and 24 with non-inflammatory diseases served as controls. MicroRNA expression levels in vitreous humor samples were measured by microarray, and differentially expressed miRNAs between sarcoidosis and other diseases were explored. Next, pathway enrichment analysis was performed to evaluate the functions of the dysregulated miRNAs, and machine learning was used to search for candidate biomarkers. RESULTS: A total of 614 upregulated miRNAs and 8 downregulated miRNAs were detected in vitreous humor of patients with ocular sarcoidosis compared with patients with unclassified uveitis and non-inflammatory diseases. Some dysregulated miRNAs were involved in the TGF-β signaling pathway. Furthermore, we identified miR-764 as the best predictor for ocular sarcoidosis using Boruta selection. CONCLUSIONS: In this study, comprehensive miRNA analysis of vitreous humor samples identified dysregulated miRNAs in ocular sarcoidosis. This study suggests new insights into molecular pathogenetic mechanisms of sarcoidosis and may provide useful information toward developing novel diagnostic biomarkers and therapeutic targets for sarcoidosis.