Introduction: COVID-19 is the most frequent cause of ARDS, generating mild, moderate and severe clinical symptoms, in the latter, patients could evolve to the need for orotracheal intubation and advanced ventilatory support; reporting mortality of 26-90% in these cases. This is why there is a need to identify risk factors and predictive tools for mortality. In this study, the diagnostic validity of the oxygen saturation and oxygenation indices as predictors of mortality prognosis in patients with severe ARDS due to COVID-19 who required invasive mechanical ventilation was determined. Material and methods: cross-sectional, analytical and retrospective study. The sample consisted of 176 medical records of critically ill patients. Bivariate and multivariate analyzes were performed, in addition to determining specificity, sensitivity, PPV, NPV, and AUC. Results: 57.37% (101/176) of the patients died, with a predominance of males 79% (80/101); with an average age of 56 years. Lactate (p=0.115) did not show a statistically significant association. Results: 57.37% (101/176) of the patients died, with a predominance of males 79% (80/101); with an average age of 56 years. Lactate (p=0.115) did not show a statistically significant association. Regarding the indices, OSI registered sensitivity and specificity of 33% and 96%, respectively; while IO had 33.6% sensitivity and 97.33% specificity; AUC 0.773 (CI: 0.677-0.868); that is, more than 77.3% probability of prognosis for mortality. Conclusion: OSI and OI are adequate predictors of mortality, with OI having better diagnostic validity. In addition, the male sex and increased glycemic values are confirmed as factors associated with the risk of developing serious conditions due to COVID-19.