The aim of this study was to investigate whether respiratory variations in carotid and aortic blood flows measured by Doppler ultrasonography could accurately predict fluid responsiveness in critically ill children. This was a prospective single-center study including mechanically ventilated children who underwent fluid replacement at the discretion of the attending physician. Response to fluid load was defined by a stroke volume increase of more than 15%. Maximum and minimum values of velocity peaks were determined over one controlled respiratory cycle before and after volume expansion. Respiratory changes in velocity peak of the carotid (∆Vpeak_Ca) and aortic (∆Vpeak_Ao) blood flows were calculated as the difference between the maximum and minimum values divided by the mean of the two values and were expressed as a percentage. A total of 30 patients were included, of which twelve (40%) were fluid responders and 18 (60%) non-responders. Before volume expansion, both ∆Vpeak_Ca and ∆Vpeak_Ao were higher in responders than in non-responders (17.1% vs 4.4%; p < .001 and 22.8% vs 6.4%; p < .001, respectively). ∆Vpeak_Ca could effectively predict fluid responsiveness (AUC 1.00, 95% CI 0.88-1.00), as well as ∆Vpeak_Ao (AUC 0.94, 95% CI 0.80-0.99). The best cutoff values were 10.6% for ∆Vpeak_Ca (sensitivity, specificity, positive predictive value and negative predictive value of 100%) and 18.2% for ∆Vpeak_Ao (sensitivity, 91.7%; specificity, 88.9%; positive predictive value, 84.6%; negative predictive value, 94.1%). Volume expansion-induced changes in stroke volume correlated with the ∆Vpeak_Ca and ∆Vpeak_Ao before volume expansion (ρ of 0.70 and 0.61, respectively; p < .001 for both). Analysis of respiratory changes in carotid and aortic blood flows are accurate methods for predicting fluid responsiveness in children under invasive mechanical ventilation.