Pulmonary complications have a significant impact on morbidity and mortality in patients after major surgery. Lung ultrasound can be used at the bed-side, and has gained widespread acceptance in the intensive care unit. We conducted a prospective study to evaluate whether lung ultrasound could be used as a predictive marker for postoperative ventilatory support in high-risk surgical patients. We included 109 patients admitted to the intensive care unit while having mechanical ventilation of the lungs following major surgery. The PaO2/FIO2 ratio was calculated on admission and an ultrasound examination performed, including: lung ('lung ultrasound score', number of consolidated lung areas); cardiac (mitral flow); and inferior vena cava imaging (diameter and respiratory variation). Respiratory outcomes included: the need for ventilation support (mechanical ventilation, non-invasive ventilation or high-flow nasal cannula oxygen therapy); acute respiratory distress syndrome; cardiogenic pulmonary oedema; and early or late pulmonary infection. Patients with a lung ultrasound score >= 10 had a lower PaO2/FIO2 ratio, and needed more postoperative ventilatory support, than patients with lung ultrasound score < 10. Twenty patients had acute respiratory distress syndrome, and 14 had cardiogenic pulmonary oedema. The presence of >= 2 areas of consolidated lung was associated with a lower PaO2/FIO2 ratio, postoperative ventilatory support, longer intensive care stay and episodes of ventilator-associated pneumonia requiring antibiotics. Our results suggest that at intensive care unit admission, lung ultrasound scoring and detection of atelectasis can predict postoperative pulmonary outcomes after major visceral surgery, and could enhance bed-side decision making.