Stroke is the second leading cause of death worldwide. Understanding of gene expression dynamics could bring new approaches in diagnostics and therapy of stroke. Small noncoding molecules termed 'microRNA' represent the most flexible network of gene expression regulators. To screen out miRNAs that are mainly regulated during reperfusion in mechanically embolized patients, and study their mechanisms of action in reperfusion injury after thrombectomy, in order to find new therapeutic targets for mechanically embolized patients. Serums from 30 patients with moderate to severe stroke after mechanical thrombectomy (MT) were collected to measure miRNA expressions. Clinical information of patients was analyze, and patients were divided into poor prognosis and good prognosis. Factors affecting prognosis was classified, and independent risk factors for poor prognosis were determined. Prognostic value of National Institutes of Health Stroke Scale (NIHSS) score on admission to patients with MT was assessed. ROC (receiver operating characteristic) curves were drawn, and Kaplan-Merier method determined whether different NIHSS scores at admission had any difference in the in-hospital survival rate of consistency index/random consistency index (CI/RI) patients treated with MT. An oxygen-glucose deprivation/reperfusion (OGD/R) cell model and an middle cerebral artery occlusion (MCAO)/reperfusion mouse model were established, in which miR-298 expression was tested. In OGD/R cells, proliferation, apoptosis, and autophagy were assessed after intervention with miR-298 and/or autophagy related gene 5 (ATG5). In MCAO mice, the infarct area was calculated, and neurological function was assessed. The relationship between miR-298 and ATG5 was explored and validated. Age, diabetes, hypertension, hemorrhage transformation, NIHSS score at admission, leukocyte, neutrophil count and neutrophil to lymphocyte ratio (NLR) level were associated with patient's prognosis. Diabetes, NIHSS score at admission, and hemorrhagic transformation were independent risk factors for predicting poor prognosis in patients treated with MT. NIHSS score on admission had a predictive value on patient's prognosis. miR-298 was upregulated in acute cerebral ischemia patients with MT (p<0.05), especially in those with poor prognosis. miR-298 was elevated in both cell and mouse models (p<0.05). Apoptosis and autophagy of cells were weakened after miR-298 knockdown, and infarction in the mouse brain tissues was reduced. ATG5 was a target of miR-298. Overexpressing ATG5 rescued miR-298-induced apoptosis and autophagy. In conclusion: regulation of miR-298 and ATG5 attenuates neuronal apoptosis and autophagy, providing a new strategy for brain injury after reperfusion in patients with MT.