This retrospective study examined changes in fasting blood glucose (FBG) levels during hospitalization and their effect on risk of death for Coronavirus disease 2019 (COVID-19) patients without previously diagnosed diabetes . A model with low- and high-stable pattern trajectories was established based on a longitudinal change in FBG levels . We analyzed FBG trajectory-associated clinical features and risk factors for death due to COVID-19 . Of the 230 enrolled patients , 44 died and 87.83% had a low-stable pattern (average FBG range : 6.63-7.54 mmol/L), and 12.17% had a high-stable pattern (average FBG range : 12.59-14.02 mmol/L). There were statistical differences in laboratory findings and case fatality between the two FBG patterns . Multivariable logistic regression analysis showed that increased neutrophil count (odds ratio [OR], 25.43; 95% confidence interval [CI]: 2.07 , 313.03), elevated direct bilirubin (OR , 5.80; 95% CI : 1.72 , 19.58), elevated creatinine (OR , 26.69; 95% CI : 5.82 , 122.29), lymphopenia (OR , 8.07; 95% CI : 2.70 , 24.14), and high-stable FBG pattern (OR , 8.79; 95% CI : 2.39 , 32.29) were independent risk factors for higher case fatality in patients with COVID-19 and hyperglycemia but no history of diabetes . FBG trajectories were significantly associated with death risk in patients with COVID-19 and no diabetes.