Objective: To investigate the relationship between blood glucose variability and prognosis in patients with massive cerebral infarction. Materials and methods: From July 2013 to December 2015, a total of 111 patients with massive cerebral infarction were enrolled in this study, divided into survival group (n = 88) and death group (n = 23). We analyzed blood glucose levels in both groups, average blood glucose (GluAve), standard deviation of blood glucose (GluSD), blood glucose coefficient of variation (GluCV), large amplitude of glycemic excursion (LAGE), mean amplitude of glycemic excursion (MAGE), mean postprandial glucose excursion (MPPGE) was calculated and the relationship between blood glucose variability and patient mortality was compared between the two groups. Results: The difference in the GluAve, SD, GluCV, LAGE, MAGE, MPPGE between the death group and the survival group was all statistically significant (P < 0.001). The Pearson test results showed that, GluSD, GluCV, LAGE, MAGE and patients' outcomes were highly correlated (Pearson r > 0.9) and had significant differences (P < 0.001), respectively. But GluAve, MPPGE were middling correlated (Pearson r < 0.5), with patients' outcomes, respectively and had significant differences (P < 0.001). Conclusions: Blood glucose variability is closely related to the death of patients with massive cerebral infarction, which can better predict the prognosis of patients than the average blood glucose level.