Abstract Background During the COVID-19 pandemic, creating tools to assess disease severity is one of the most important aspects of reducing the burden on emergency departments. Lung ultrasound has a high accuracy for the diagnosis of pulmonary diseases; however, there are few prospective studies demonstrating that lung ultrasound can predict outcomes in COVID-19 patients. We hypothesized that lung ultrasound score (LUS) at hospital admission could predict outcomes of COVID-19 patients. This is a prospective cohort study conducted from 14 March through 6 May 2020 in the emergency department (ED) of an urban, academic, level I trauma center. Patients aged 18 years and older and admitted to the ED with confirmed COVID-19 were considered eligible. Emergency physicians performed lung ultrasounds and calculated LUS, which was tested for correlation with outcomes. This protocol was approved by the local Ethics Committee number 3.990.817 (CAAE: 30417520.0.0000.0068). Results The primary endpoint was death from any cause. The secondary endpoints were ICU admission and endotracheal intubation for respiratory failure. Among 180 patients with confirmed COVID-19 who were enrolled (mean age, 60 years; 105 male), the average LUS was 18.7 ± 6.8. LUS correlated with findings from chest CT and could predict the estimated extent of parenchymal involvement (mean LUS with < 50% involvement on chest CT, 15 ± 6.7 vs. 21 ± 6.0 with > 50% involvement, p < 0.001), death (AUC 0.72, OR 1.13, 95% CI 1.07 to 1.21; p < 0.001), endotracheal intubation (AUC 0.76, OR 1.17, 95% CI 1.09 to 1.26; p < 0.001), and ICU admission (AUC: 0.71, OR 1.14, 95% CI 1.07 to 1.21; p < 0.001). Conclusions In COVID-19 patients admitted in ED, LUS was a good predictor of death, ICU admission, and endotracheal intubation.