microRNAs (miRNA) are recognized as regulators of gene expression during development and cell differentiation as well as biomarkers of disease. Development of rapid and sensitive miRNA profiling methods is essential for evaluating the pattern of miRNA expression that varies across normal and diseased states. The ability to identify miRNA expression patterns is limited to cumbersome assays that often lack sensitivity and specificity to distinguish between different miRNA families and members. We evaluated a surface-enhanced Raman scattering (SERS) platform for detection and classification of miRNAs. The strength of the SERS-based sensor is its sensitivity to detect extremely low levels of analyte and specificity to provide the molecular fingerprint of the analyte. We show that the SERS spectra of related and unrelated miRNAs can be detected in near-real time, that detection is sequence dependent, and that SERS spectra can be used to classify miRNA patterns with high accuracy. © 2008 Elsevier B.V. All rights reserved.