Lung ultrasound may be useful for prognostication of acute lung disease. To assess whether the lung ultrasound score is associated with the severity of lung disease and may predict prolonged invasive mechanical ventilation in critically ill children. Prospective observational multicenter study in children aged 1 month to 18 years who required respiratory support in the intensive care unit. Children with chronic parenchymal lung disease were excluded. The lung ultrasound score was obtained at 12 hours and 48-72 hours from admission. Prolonged invasive mechanical ventilation was defined as >7 consecutive days. Correlation of the lung ultrasound score with oxygenation as well as its prognostic accuracy for prolonged invasive mechanical ventilation were investigated. 538 children were included and 62 (11.5%) required prolonged mechanical ventilation. In these subjects, the lung ultrasound score was higher at 12 [24 (19-26) vs. 8 (3-14); p<0.001] and 48-72 hours [16 (10.5-22.5) vs. 6 (3-11) vs; p<0.001]. At 12 hours the lung ultrasound score correlated with oxygenation index [R2= 0.435 (95% CI: 0.293-0.566), rho coefficient -0.705, p<0.001] and oxygen saturation index [R2 0.499 (95% CI: 0.370-0.613), rho coefficient 0.651, p<0.001p<0.001]. To predict prolonged invasive mechanical ventilation, the lung ultrasound score at 12 hours had a good accuracy [AUROC=0.87 (95% CI: 0.81-0.93)] while its use in a multivariable model had an excellent accuracy both in derivation [AUROC=0.92 (95% CI: 0.89-0.95)] and internal validation [AUROC=0.91 (95% CI: 0.90-0.92)]. In critically ill children, the lung ultrasound score early after admission may predict prolonged invasive mechanical ventilation.