Introduction: SARS-CoV-2 infection can induce disturbances of several biochemical parameters especially in patients with severe or critical forms of the disease. The aim of this study is to evaluate the predictive markers of the severity of COVID-19.Materials and methods: Retrospective descriptive study, involving 129 COVID-19 positive patients. Patients were classified into two cohorts: severe (n=103) and non-severe (n=26) cases. The ROC curve and the associated area under the curve determination were used to determine the predictors of disease severity.Results: The comparison between the two groups shows that the values of AST, GGT, LDH, CPK, CRP, Ferritin, PCT, Urea, Magnesium, Troponin I and BNP were significantly higher in patients with severe forms of COVID-19 compared to those with non-severe forms. While albumin, Chlorine and Calcium were significantly decreased in severe patients. ROC curve analysis showed that Albumin (AUC = 0.779), LDH (AUC = 0.798) and PCT (AUC = 0.732) have a medium predictive value for the severity of COVID-19. While the prognostic value were good for BNP (AUC = 0.854), CRP (AUC = 0.845) and high sensitivity Troponin I (AUC = 0.812).Conclusion: The identification of biological factors predictive of the severity and/or mortality of COVID-19 infection will allow the development of predictive models of disease severity and their complementary uses in clinical practice for risk stratification and for early and adapted management to minimize the mortality rate and for rational allocation of medical resources.