This study aimed to evaluate the association between the glucose-to-lymphocyte ratio (GLR) and all-cause mortality in intensive care unit (ICU) patients with Non-traumatic cerebral hemorrhage. This is a retrospective cohort study. Baseline data and in-hospital prognosis from patients with non-traumatic cerebral hemorrhage admitted to the intensive care unit. Multivariate COX regression analysis was applied and adjusted hazard ratios (HR) and 95% predictive values with confidence intervals (CI) were calculated. Survival curves for the two groups of cases were plotted using K-M curves, and subgroup analyses were performed in one step. Using restricted cubic spline curves, we analyzed the potential linear relationship between GLR and outcome indicators. In the Medical Information Mart for Intensive Care IV (MIMIC-IV database), we extracted 3,783 patients with nontraumatic intracerebral hemorrhage, and 1,806 patients were finally enrolled in the study after exclusion of missing values and patients with a short hospital stay. The overall ICU mortality rate was 8.2% (148/1806) and the in-hospital mortality rate was 12.5% (225/1806). The use of curve fitting yielded a significant linear relationship between GLR and both ICU mortality and in-hospital mortality. It also suggested a reference point at GLR=3.9. These patients were categorized into high and low subgroups based on the median value of their GLR (GLR = 3.9). Model comparisons based on multivariate COX regression analysis showed that in-hospital mortality was higher in the high GLR group after adjusting for all confounders (HR = 1.31, 95% CI: 1.04-1.47), while the ICU mortality in the high GLR group was (HR = 1.73, 95% CI: 1.18-2.52). Stratified analyses based on age, gender, race, GCS, BMI, and disease type showed stable correlations between the high GLR group and in-hospital and ICU mortality. Based on our retrospective analysis, it is known that as the GLR increased, the in-hospital mortality rate and ICU mortality rate of patients with nontraumatic cerebral hemorrhage also increased progressively in the United States in a clear linear relationship. However, further studies are needed to confirm these findings.