Objective To investigate the relationship between the severity of COVID-19 and the changes of lactate dehydrogenase (LDH-L), creatine kinase (CK), creatine kinase MB isoenzyme (CK-MB), α-hydroxybutyrate dehydrogenase (α-HBDH) in vitro. Methods Retrospective analysis of serum LDH-L, CK, CK-MB, α-HBDB and clinical characteristics of 58 patients with confirmed COVID-19 admitted to the Fourth People's Hospital of Nanning, Guangxi from January 18th to March 2nd in 2020. According to COVID-19 diagnosis and treatment program (trial 7th edition), the patients were classified into two groups: common type (including mild type) group, severe type and critical type group. with following information obtained, including heart rate, LDH-L, CK, CK-MB, α-HBDH biochemical indexes, basic information, and clinical characteristics. A logistic regression model was built to analyze the independent risk factors for COVID-19. Results There were 51 cases of mild and common type, 7 cases of heavy and critical type (4 cases of heavy type and 3 cases of critical type). A single-factor analysis showed that the gender ratio of control group was similar between the common type (including mild type) and the heavy and critical type (86.21% vs 89.66%,13.79% vs 10.34%, χ2=0.162, P=0.687), The common type and mild type are mainly young adults, while the heavy and critical types are elders [(41.48±17.49)years old vs (68.86±17.49)years old, t=-9.936, P<0.001]; Fewer complications of cardiovascular and cerebrovascular diseases and diabetes were present in common and mild cases than that in heavy and critical types with significant difference by Fisher precision test (23.53 vs 85.71%, P<0.001).However, no significant difference was found in heart rate (t=-0.180, P=0.858), CK-MB (t=1.531, P=0.131), and CK (t=-1.449, P=0.197), but not for LDH [(200.67±49.44)U/L vs (385.02±86.33)U/L, t=-5.527, P=0.001], and α-HBDH [(149.43±34.85)U/L vs (289.15±56.70)U/L, t=-6.357, P<0.001]. The suspicious risk factors with significance selected by single factor analysis and clinical important discriminators were analyzed by collinearity analysis and spearman correlation analysis to select the suspicious risk factors for multivariate logistic regression analysis, which showed that LDH-L was an independent risk factor for the occurrence of COVID-19 with odds ratio of 1.037 (95% CI=1.010-1.065, P=0.006)and accuracy in prediction of 93.1%. Conclusion LDH-L is an independent risk factor for assessing the severity of the COVID-19, which can guide early effective treatment in the clinical scenarios.