MicroRNAs have been reported as biomarkers for various diseases, including cerebral atherosclerosis (AS). In this study, whether serum microRNA-137 (miR-137) could be used as a biomarker for diagnosing cerebral AS and predicting cerebrovascular event was investigated. Quantitative real-time PCR was used to measure the expression of miR-137 in serum. Logistic analysis was used to evaluate the risk factors for the occurrence of cerebral AS, and receiver operating characteristic curves were used to estimate the diagnostic value of miR-137 and other risk factors for AS occurrence. Furthermore, the prognostic value of miR-137 for patients with AS was estimated using Kaplan-Meier survival analysis and Cox regression analysis. The results indicated that serum miR-137 levels were decreased in patients with cerebral AS. The expression of miR-137 was negatively correlated with total cholesterol and low-density lipoprotein cholesterol levels in patients with cerebral AS. The levels of miR-137, total cholesterol, low-density lipoprotein cholesterol, and hypersensitivity C response protein may serve as risk factors for the occurrence of cerebral AS, and miR-137 had diagnostic value for AS screening. Cerebral AS patients with positive cerebrovascular events have low miR-137 expression. Patients with high miR-137 expression had a lower incidence of cerebrovascular adverse events (log-rank P = 0.013), and miR-137 was an independent prognostic marker for the prediction of cerebrovascular event occurrence in patients with cerebral AS. In conclusions, our findings indicate that serum miR-137 levels are decreased in patients with cerebral AS and may be a new biomarker for diagnosing cerebral AS and predicting cerebrovascular events.