Inducing general anesthesia (GA) in children can considerably affect blood pressure, and the rate of severe critical events owing to this remains high. Cerebrovascular autoregulation (CAR) protects the brain against blood-flow-related injury. Impaired CAR may contribute to the risk of cerebral hypoxic-ischemic or hyperemic injury. However, blood pressure limits of autoregulation (LAR) in infants and children are unclear. In this pilot study CAR was monitored prospectively in 20 patients aged <4 years receiving elective surgery under GA. Cardiac- or neurosurgical procedures were excluded. The possibility of calculating the CAR index hemoglobin volume index (HVx), by correlating near-infrared-spectroscopy (NIRS)-derived relative cerebral tissue hemoglobin and invasive mean arterial blood pressure (MAP) was determined. Optimal MAP (MAPopt), LAR, and the proportion of time with a MAP outside LAR were determined. The mean patient age was 14 ± 10 months. MAPopt could be determined in 19 of 20 patients, with an average of 62 ± 12 mmHg. The required time for a first MAPopt depended on the extent of spontaneous MAP fluctuations. The actual MAP was outside the LAR in 30% ± 24% of the measuring time. MAPopt significantly differed among patients with similar demographics. The CAR range averaged 19 ± 6 mmHg. Using weight-adjusted blood pressure recommendations or regional cerebral tissue saturation, only a fraction of the phases with inadequate MAP could be identified. Non-invasive CAR monitoring using NIRS-derived HVx in infants, toddlers, and children receiving elective surgery under GA was reliable and provided robust data in this pilot study. Using a CAR-driven approach, individual MAPopt could be determined intraoperatively. The intensity of blood pressure fluctuations influences the initial measuring time. MAPopt may differ considerably from recommendations in the literature, and the MAP range within LAR in children may be smaller than that in adults. The necessity of manual artifact elimination represents a limitation. Larger prospective and multicenter cohort studies are necessary to confirm the feasibility of CAR-driven MAP management in children receiving major surgery under GA and to enable an interventional trial design using MAPopt as a target.