Background Traumatic brain injury (TBI) is a growing public health concern that can be complicated with an acute stress response. This response may be assessed by monitoring blood glucose levels but this is not routine in remote settings. There is a paucity of data on the prevalence of hyperglycemia and variables associated with mortality after severe TBI in Uganda. Objective We aimed to determine the prevalence of hyperglycemia in patients with severe TBI and variables associated with 30-day mortality at Mulago National Referral Hospital in Uganda. Methods We consecutively enrolled a cohort 99 patients patients with severe TBI. Serum glucose levels were measured at admission and after 24 h. Other study variables included: mechanism of injury, CT findings, location and size of hematoma, and socio-demographics. The main outcome was mortality after 30 days of management and this was compared in patients with hyperglycemia more than 11.1 mmol/L to those without. Results Most patients (92.9%) were male aged 18–30 years (47%). Road Traffic Collisions were the most common cause of severe TBI (64.7%) followed by assault (17.1%) and falls (8.1%). Nearly one in six patients were admitted with hyperglycemia more than 11.1 mmol/L. The mortality rate in severe TBI patients with hyperglycemia was 68.8% (OR 1.47; 95% CI [0.236–9.153]; P = 0.063) against 43.7% in those without hyperglycemia. The presence of hypothermia (OR 10.17; 95% CI [1.574–65.669]; P = 0.015) and convulsions (OR 5.64; 95% CI [1.541–19.554]; P = 0.009) were significant predictors of mortality. Conclusion Hypothermia and convulsions at admission were major predictors of mortality in severe TBI. Early hyperglycemia following severe TBI appears to occur with a tendency towards high mortality. These findings justify routine glucose monitoring and could form the basis for establishing a blood sugar control protocol for such patients in remote settings.