Hyperglycemia commonly occurs in severe cases with COVID-19. In this study, we explored the associations between admission fasting plasma glucose (FPG) and 28-day mortality in COVID-19 patients. In this single centre retrospective study, 263 adult patients with COVID-19 were included. Demographic and clinical information were collected and compared between patients with and without diabetes. Cox regression analyses were used to investigate the risk factors of 28-day mortality in hospitals. Of 263 patients, 161 (61.2%) were male, 62 (25.6%) had a known history of diabetes, and 135 (51.3%) experienced elevated FPG (>7.0 mmol/L) at hospital admission. The median FPG in patients with diabetes was much higher than in patients without diabetes (12.79 vs. 6.47 mmol/L). Patients with diabetes had higher neutrophil count and D-dimer, less lymphocyte count, lower albumin level, and more fatal complications. Multivariable Cox regression analyses showed that age (per 10-year increase) (hazard ratio [HR], 1.41; 95% confidence interval [CI], 1.13-1.74), admission FPG between 7.0 and 11.0 and >= 11.1 mmol/L (HR, 1.90; 95% CI, 1.11-3.25 and HR, 2.09; 95% CI, 1.21-3.64, respectively), chronic obstructive pulmonary disease (HR, 2.89; 95% CI, 1.31-6.39), and cardiac injury (HR, 2.14; 95% CI, 1.33-3.47) were independent predictors of 28-day mortality in COVID-19 patients. Hyperglycemia on admission predicted worse outcome in hospitalized patients with COVID-19. Intensive monitoring and optimal glycemic control may improve the prognosis of COVID-19 patients.