Objective To analyze the predictive factors for severe coronavirus disease 2019 (COVID-19) pneumonia, so as to provide clin. basis for early clin. intervention and reduction in severe cases and mortality. Methods The epidemiol. and demog. data, clin. manifestations, and laboratory and imaging data of 280 patients with confirmed COVID-19 were retrospectively analyzed. The subjects were divided into two groups, non-severe group (ordinary patients) and severe group (including severe and critical patients). Logistic regression anal. was used to explore the risk factors of severe cases. Results Among the 280 patients, 130 (46.4%) were male and 150 (53.6%) were female; and the median age was 49.5 years (35.0-62.0 years). Common symptoms include fever (70.7%), cough (56.4%), expectoration (23.9%) and sore throat (13.2%). Among the patients, 225 (80.4%) were non-severe cases and 55 (19.6%) were severe cases. Compared with severe patients, the age of non-severe patients was significantly lower (45.0 years vs. 61.00, P<0.001). The incidence of complications in severe patients was significantly higher than that in non-severe patients (58.2% vs. 28.0%, P<0.001). Compared with non-severe patients, in severe patients, a the levels of neutrophils, neutrophil/lymphocyte ratio, alanine aminotransferase, aspartate aminotransferase, C-reactive protein, procalcitonin, lactate dehydrogenase, creatine kinase and D-dimer in severe patients were significantly higher (all P<0.05); the levels of lymphocytes, erythrocytes and platelets were significantly lower (P<0.05); bilateral lung lesion were significantly obvious (P<0.001); and the pos. rate of novel coronavirus nucleic acid in peripheral blood was significantly higher (36.2% vs. 3.1%, χ2=41.478, P=0.000). Multivariate logistic regression anal. showed that age, neutrophil/lymphocyte ratio, aspartate aminotransferase, and pos. blood nucleic acid were independent risk factors for exacerbation of patients with COVID-19. The risk of progression to severe disease in patients with pos. blood nucleic acid was significantly increased (OR=21.666, 95%CI: 5.143-91.268). ROC curve anal. showed that the areas under the ROC curves (AUC) of age, NLR, AST, pos. predictive value of serum nucleic acid and predictive value of logistic regression model for predicting severe cases of COVID-19 patients were 0.728, 0.798, 0.737, 0.665 and 0.890, resp. (all P<0.01). Conclusion Patients with older age and more complications are more likely to develop severe diseases. Age, neutrophil/lymphocyte ratio, aspartate aminotransferase and pos. serum nucleic acid are independent risk factors for severe COVID-19.