Background: Prior studies reported that 5~32% COVID-19 patients were critically ill, a situation that poses great challenge for the management of the patients and ICU resources. We aim to identify independent risk factors to serve as prediction markers for critical illness MESHD of SARS-CoV-2 infection MESHD. Methods : Fifty-two critical and 200 non-critical SARS-CoV-2 nucleic acid positive patients hospitalized in 15 hospitals outside Wuhan from January 19 to March 6, 2020 were enrolled in this study. Multivariable logistic regression and LASSO logistic regression were performed to identify independent risk factors for critical illness MESHD.Results: Age TRANS older than 60 years, dyspnea MESHD dyspnea HP, respiratory rate > 24 breaths per min, leukocytosis MESHD leukocytosis HP >9.5 X109/L, neutrophilia HP >6.3 X109/L, lymphopenia MESHD lymphopenia HP <1.1 X109/L, neutrophil-to-lymphocyte ratio >3.53, fibrinogen >4g/L, d-dimer >0.55 µg/mL, blood SERO urea nitrogen >7.1 mM, elevated aspartate transaminase, elevated alanine aminotransferase, total bilirubin >21 µM, and Sequential Organ Failure Assessment (SOFA) score ≥2 were identified as risk factors for critical illness MESHD. LASSO logistic regression identified the best combination of risk factors as SOFA score, age TRANS, dyspnea MESHD dyspnea HP, and leukocytosis MESHD leukocytosis HP. The Area Under the Receiver-Operator Curve values for the risk factors in predicting critical illness MESHD were 0.921 for SOFA score, 0.776 for age TRANS, 0.764 for dyspnea MESHD dyspnea HP, 0.658 for leukocytosis MESHD leukocytosis HP, and 0.960 for the combination of the four risk factors.Conclusions: Our findings advocate the use of risk factors SOFA score ≥2, age TRANS >60, dyspnea MESHD dyspnea HP and leukocytosis MESHD leukocytosis HP >9.5 X109/L on admission, alone or in combination, to determine the optimal management of the patients and health care resources.