In the last years, biomarkers of infection, such as the soluble urokinase plasminogen activator receptor (suPAR), have been extensively studied as potential diagnostic and prognostic biomarkers in the intensive care unit (ICU). In this study, we investigated whether this biomarker can be used in COVID-19 and non-COVID-19 septic patients for mortality prediction. Serum suPAR levels were measured in 79 non-COVID-19 critically ill patients upon sepsis (within 6 h), and on admission in 95 COVID-19 patients (66 critical and 29 moderate/severe). The non-COVID-19 septic patients were matched for age, sex, and disease severity, while the site of infection was the respiratory system. On admission, COVID-19 patients presented with higher suPAR levels, compared to non-COVID-19 septic patients (p < 0.01). More importantly, suPAR measured upon sepsis could not differentiate survivors from non-survivors (p > 0.05), as opposed to suPAR measured on admission in COVID-19 survivors and non-survivors (p < 0.0001). By the generated ROC curve, the prognostic value of suPAR in COVID-19 was 0.81, at a cut-off value of 6.3 ng/mL (p < 0.0001). suPAR measured early (within 24 h) after hospital admission seems like a specific and sensitive mortality risk predictor in COVID-19 patients. On the contrary, suPAR measured at sepsis diagnosis in non-COVID-19 critically ill patients, does not seem to be a prognostic factor of mortality.