Immunoglobulin A nephropathy (IgAN) is a worldwide disease characterized by the presence of galactose-deficient IgA1 deposits in the glomerular mesangium. A kidney biopsy for diagnosis is required. Here, we measured two miRNAs (let-7b and miR-148b), previously identified as regulators of the O-glycosylation process of IgA1, in serum samples from patients with IgAN and healthy blood donors (controls) recruited in an international multicenter study. Two predictive models, based on these miRNAs, were developed and the diagnostic accuracy of the combined biomarkers was assessed by the area under the receiver operating characteristic (ROC) curve (AUC) carried out in three steps. In a training study, the combined miRNAs were able to discriminate between 100 patients with IgAN and 119 controls (AUC, 0.82). A validation study confirmed the model in an independent cohort of 145 patients with IgAN and 64 controls (AUC, 0.78). Finally, in a test study, the combined biomarkers were able to discriminate patients with IgAN from 105 patients affected by other forms of primary glomerulonephritis, supporting the specificity (AUC, 0.76). Using the same study design, we also performed two subgroup analyses (one for Caucasians and one for East Asians) and found that race-specific models were the best fit to distinguish IgAN patients from controls. Thus, serum levels of the combined miRNA biomarker, let-7b and miR-148b, appears to be a novel, reliable, and noninvasive test to predict the probability of having IgAN.