In pediatric anesthesia, deviations from normothermia can lead to many complications, with infants and young children at the highest risk. A measurement method for core temperature must be clinically accurate, precise and should be minimally invasive. Zero-heat-flux (ZHF) temperature measurements have been evaluated in several studies in adults. We assessed the agreement between the 3M Bair Hugger™ temperature measurement sensor (TZHF) and esophageal temperature (TEso) in children up to and including 6 years undergoing surgery with general anesthesia. Data were recorded in 5 min-intervals. We investigated the accuracy of the ZHF sensor overall and in subgroups of different age, ASA classification, and temperature ranges by Bland–Altman comparisons of differences with multiple measurements. Change over time was assessed by a linear mixed model regression. Data were collected in 100 children with a median (1st–3rd quartile) age of 1.7 (1–3.9) years resulting in 1254 data pairs. Compared to TEso (range from 35.3 to 39.3 °C; median 37.2 °C), TZHF resulted in a mean bias of +0.26 °C (95% confidence interval +0.22 to +0.29 °C; 95% limits of agreement −0.11 to +0.62 °C). Lin’s concordance correlation coefficient was 0.89. There was no significant or relevant change of temperature over time (0.006 °C per hour measurement interval, p = 0.199) and no relevant differences in the subgroups. Due to the mean bias of +0.26 °C in TZHF, the risk of hypothermia may be underestimated, while the risk of hyperthermia may be overestimated. Nevertheless, because of its high precision, we consider ZHF valuable for intraoperative temperature monitoring in children and infants.