Background: The outbreak of COVID-19 has a significant impact on the health of people around the world. In the clinical condition of COVID-19, the condition of critical cases changes rapidly with a high mortality rate. Therefore, early prediction of disease severity and active intervention play an important role in the prognosis of severe patients.Methods: All the patients with COVID-19 in Taizhou city were retrospectively included and segregated into the non-severe and severe group according to the severity of the disease. The clinical manifestations, laboratory examination results, and imaging findings of the 2 groups were analysed for comparing the differences between the 2 groups. Univariate and multivariate logistic regression were used for screening the factors that could predict the disease, and the nomogram was constructed.Results: A total of 143 laboratory-confirmed cases were included in the study, including 110 non-severe patients and 33 severe patients. The median age of patients was 47 years (range, 4–86 years). Fever (73.4%) and cough (63.6%) were the most common initial clinical symptoms. By using the method of multivariate logistic regression, the variables to construct nomogram include age (OR: 1.052, 95% CI: 1.020–1.086, P = 0.001), body temperature (OR: 2.252, 95% CI: 1.139–4.450, P = 0.020), lymphocyte count (OR: 1.128, 95% CI: 1.000–1.272, P = 0.049), ADA (OR: 1.163, 95% CI: 1.023–1.323, P = 0.021), PaO2 (OR: 0.972, 95% CI: 0.953–0.992, P = 0.007), IL-10 (OR: 1.184, 95% CI: 1.037–1.351, P = 0.012), and bronchiectasis (OR: 3.818, 95% CI: 1.694–8.605, P = 0.001). The AUC of the established nomogram was 0.877.Conclusions: This study established a stable nomogram for predicting the severity of COVID-19, and the clinicians can use the established nomogram for predicting the severity of newly diagnosed COVID-19 patients and to conduct active intervention for minimising the mortality rate and improving the prognosis of severe patients.