To investigate the brain function and prognosis of patients with acute and severe cerebrovascular disease by long-term electroencephalogram (EEG). Methods: Sixty-two patients with critical acute cerebrovascular disease and Glasgow coma scores (GCS) of <8 points were enrolled within 72 hours of onset. Patients were monitored by long-term EEG, and EEG grading and GCS scoring were performed. Relationships between EEG grades and GCS scores and prognosis of patients were analyzed, and long-term EEG grades and GCS scores were calculated and compared to predict patient prognosis. Results: For patients in this study, long-term EEG grade was negatively correlated with GCS score (r=－ 0.739, P<0.001); long-term EEG grade was positively correlated with patient prognosis (r=0.387, P=0.002). The accuracy of prognosis was 67.7% with long-term EEG and 62.9% with GCS score. The predictive accuracy of EEG grade on patient prognosis was greater than that of GCS score (P<0.001). Conclusion: Long-term EEG grade is significantly associated with the prognosis of critically ill patients with severe cerebrovascular disease. The accuracy of EEG grades is higher than that of GSC scores in the prognostic evaluation of critical acute cerebrovascular disease.