Serum mid-regional proadrenomedullin (MR-proADM) has emerged as a marker of organ failure (mainly lungs and kidneys) and poor prognosis in patients admitted to intensive care (IC); some reports also suggest it and other markers, such as Krebs von den Lungen-6 (KL-6) and interleukin-6 (IL-6), as a prognostic biomarker of COVID-19. The aim of the study was to evaluate the performance MR-proADM in hospitalized COVID-19 patients for predicting in-hospital mortality and need for non-invasive or invasive respiratory support. We enrolled 74 patients hospitalized in the COVID Unit of Siena Hospital from March to May 2020, for whom serum samples were available on admission for assay of MR-proADM, KL-6 and IL-6. Demographic data, comorbidities, medical history and clinical laboratory data on days 1-3 of admission and Simplified Acute Physiology Score and Simplified Organ Failure Assessment scores calculated at day 1 were collected retrospectively, as well as mortality and IC admission data. 12 patients died in hospital (16%) and 14 patients were admitted to IC (19%). Serum concentrations of MR-proADM on admission and on day 1 were higher among non-survivors than among survivors (p = 0.015 and p = 0.045, respectively), while those on day 3 were not significantly different. Patients needing respiratory support had higher MR-proADM concentrations on admission than the others (p = 0.046), and those requiring invasive mechanical ventilation had higher MR-proADM on day 1 (p = 0.017). Serum concentrations of KL-6 and IL-6 were significantly higher in non-survivors (p = 0.03 and p = 0.004, respectively). ROC curve analysis showed that serum MR-proADM on day 1 had the best accuracy in predicting death and/or IC admission (AUC = 0.9583, p = 0.0006); the combination of all three biomarkers further improved the accuracy of prediction of death or IC admission (AUC = 0.9793; p = 0.00004). Our data sustain the potential of serum MR-proADM as a reliable prognostic biomarker of hospitalized COVID-19 patients and confirms the utility of the three markers in the management and risk stratification of hospitalized patients. The markers are collected mini-invasively and are quick to analyze and cost-effective.