MicroRNAs (miRNAs) are small non-coding RNA molecules that are involved in a variety of cellular functions. Single nucleotide polymorphisms (SNPs) have been identified in mature miRNAs (mmSNPs), some of which have been linked to cancer risk; however, it is unclear which mmSNPs contribute to the susceptibility to thyroid tumors. In the present study, we examined the influence of selected mmSNPs on the risk of thyroid tumor. After systematic in silico screening, seventeen mmSNPs were identified and genotyped in a Chinese population including 828 patients with papillary thyroid cancer (PTC), 488 patients with benign thyroid tumor (BN), and 1038 cancer-free controls. Multivariate logistic regression analyses were conducted to evaluate the association of SNP genotypes and alleles with the risk of developing PTC and BN. Three SNPs (rs67106263 in mir-3144, GA versus GG, OR = 1.35, 1.09-1.68; rs4919510 in mir-608, CC versus GG/GC, OR = 0.76, 0.60-0.97; and rs79402775 in mir-933, AA versus GG/GA, OR = 1.76, 1.00-3.12) were associated with PTC risk. A combined effect of unfavorable genotypes was observed to give increased PTC risk in a dose-dependent manner. In addition, three SNPs (rs10061133 in mir-449b, rs79402775 in mir-933 and rs4919510 in mir-608) showed at least borderline correlations with the risk of BN. False-positive report probability was assessed for significant findings. The rs67106263 SNP was associated with the expression level of mir-3144 in thyroid tissue. These results indicate that mmSNPs may contribute to genetic susceptibility to thyroid tumors. Large validation and functional studies are required to further explore the role of mmSNPs in carcinogenesis.