We have developed a rapid, sensitive, and quantitative method for identification of microRNA (miRNA) sequences in multicomponent mixtures using surface-enhanced Raman spectroscopy (SERS). The method uses Ag nanorod array substrates prepared by oblique angle vapor deposition as the SERS platform. We show that Ag nanorod-based SERS spectra are uniquely characteristic for each miRNA sequence studied, and that the spectral reproducibility is sufficient for quantitative analysis of miRNA profiles in multicomponent mixtures using partial least squares (PLS) regression analysis. This method was applied to individual sample mixtures consisting of two, three, and five miRNAs. Separate PLS models were generated for the two-, three-, and five-component mixtures from >150 calibration spectra covering a concentration range of 6 to 150 microM for each miRNA. The PLS models were externally validated with independent test samples resulting in root mean square errors of prediction (RMSEP) of 7.4, <7.4, and <10 microM for the two-, three-, and five-component models, respectively. These results demonstrate the applicability of SERS for quantitative detection and profiling of miRNAs and suggest that SERS may prove to be a novel, label-free method for identification of disease biomarkers.