To explore the clinical significance of seven diabetes-related serum microRNAs (miR-9, miR-29a, miR-30d, miR34a, miR-124a, miR146a and miR375) during the pathogenesis of type 2 diabetes (T2D), 56 subjects were recruited to this study: 18 cases of newly diagnosed T2D (n-T2D) patients, 19 cases of pre-diabetes individuals (impaired glucose tolerance [IGT] and/or impaired fasting glucose [IFG]) and 19 cases of T2D-susceptible individuals with normal glucose tolerance (s-NGT). Serum miRNAs were determined by real-time RT-PCR. Expression levels of single miRNAs and the expression signatures of miR-NAs as a panel were analysed among the three groups. In n-T2D, all 7 miRNAs were significantly up-regulated compared with s-NGT and five were significantly up-regulated compared with pre-diabetes, while miRNA expression was not significantly different between s-NGT and pre-diabetes. By Canonical discriminant analysis, 70.6% of n-T2D subjects (12/17) were recognized by canonical discriminant function, while s-NGT and pre-diabetes subjects could not be discriminated from each other. Similar results were found in Hierarchical Clustering analysis based on the expression levels of all seven miRNAs. In different statistical analysis, miR-34a always showed the most significant differences. We conclude that the expression levels of seven diabetes-related miRNAs in serum were significantly elevated in n-T2D compared with prediabetes and/or s-NGT, and the latter two groups featured similar expression patterns of these miRNAs, suggesting that during the pathogenesis of T2D, the peripheral diabetes-related miRNAs have not changed significantly from s-NGT at pre-diabetic stage.