Background: MicroRNAs (miRNAs) have been identified as promising novel biomarkers for cancer diagnosis and prognosis, especially for hepatocellular carcinoma (HCC). Nowadays, the expression level of miR-21 in serum samples is a diagnostic indicator for HCC diagnosis. Thus, the quantitative determination of miRNA concentra-tion is of significance in clinical practice. It is particularly important to establish an analytical detection method for miR-21 in patient serum. Methods: The signal readout for miR-21 was based on the mass response of a repor-ter peptide using an isotope dilution mass spectrometry (MS) method in this work. To be more specific, miR-21 was biotinylated before being coupled with streptavidin (SA) agarose and then hybridized with a newly synthe-sized DNA-peptide probe. The release and purification of the sample was based on the method including trypsin digestion, solid-phase extraction, and drying, and the detection of the reporter peptide was carried out by UHPLC/MS/MS. The miR-21 in the corresponding samples was quantified by the ratio of the chromatographic peak area of the redissolved polypeptide to that of the isotope-labeled polypeptide. Additionally, within the cali-bration range, the performance of the method (including precision, accuracy, linearity, and recovery) was eval-uated. Results: The concentration of miR-21 was determined using the ratio of relative peak area of stable isotope-labeled internal standard and reporter peptide, yielding a linear range of 0.1 similar to 30.0 nM (y = 0.0818x + 0.7554, R2 = 0.9586, P < 0.01). The limit of detection (LOD) for miR-21 was 10 pM. For y5, the recoveries (n = 3) were 91.36 +/- 2.19%, 93.64 +/- 3.55%, and 96.04 +/- 2.02% for the levels of three miR-21 samples including RL, RM, and RH, respectively. Conclusions: Overall, this research provides a novel analytical approach for quan-titative detection of miRNAs in clinical serum samples.