Background COVID-19 is a public health emergency that is spreading worldwide and seriously affecting the global economy. Data on the effectiveness and safety of the use of methylprednisolone for patients with severe COVID-19 remain limited. Methods In this retrospective study, epidemiological, clinical, laboratory, treatment and outcomes data of hospitalized patients with severe COVID-19 in Zhongnan Hospital of Wuhan University from January 1 to 7 March 2020, were collected. Binary logistic regression model was used to analyse risk factors for disease progression from severe COVID-19 illness to critical illness. The effectiveness and safety of the use of methylprednisolone for patients with severe COVID-19 disease were evaluated. Results The results of the multivariate analysis from 175 patients with severe COVID-19 indicate that the use of methylprednisolone was a protective factor against disease progression from severe to critical illness(P < .001; OR: 0.054 95% CI: 0.017-0.173). Among patients with severe COVID-19 aged < 65 years, both the proportion of patients who progressed to critical illness (42.2% vs 90.0%,P = .000) and the mortality(6.7% vs 30.0%,P = .002) were lower for patients in methylprednisolone group, compared with those in the non-methylprednisolone group, whereas no statistical differences between the methylprednisolone group and the non-methylprednisolone group were found among patients with COVID-19 older than 65 years. Moreover, both the levels of CD4(+)T lymphocyte counts (646 vs 463/mu L,P = .007) and IL-6 (241.9 vs 82.8 pg/mL,P = .025) were higher among patients with severe COVID-19 aged < 65 years, compared with those patients >= 65 years old. Conclusion Data from the limited sample showed that the early use of low or medium doses of methylprednisolone has a positive effect for patients with severe COVID-19 younger than 65 years old, and excessive immune response and cytokine storm may be some of the reasons for the effectiveness.