Purpose The objective of our study was to evaluate the impact of an automatic tube potential selection (ATPS) on the delivered dose and image quality in unenhanced head computed tomography (CT) scans of infants. Materials and methods Unenhanced head CT scans were acquired before and after the introduction of an ATPS in full automatic mode in two groups of 20 patients under one year of age. The delivered dose (CDTIvol) as the quantitative (contrast-to-noise ratio) and qualitative (based on the European CT criteria) image quality were compared on the supra- and infratentorial regions by three senior pediatric radiologists. Mann–Whitney and Fisher exact tests were performed. An interobserver Fleiss's kappa agreement was calculated for each criterion. Results The use of an ATPS allowed a significant reduction in the delivered dose (−21%, p = 0.0005) with no significant difference of the contrast-to-noise ratio in supra- (−5%, p = 0.21) and infratentorial regions (+16%, p = 0.96). In all cases, dose reduction was obtained with the same value of 100 kV. It maintained a good qualitative image quality (e.g., differentiation between gray and white matter in supra-tentorial region: p = 0.470). The interobserver Fleiss's kappa agreements were good to excellent. Conclusion ATPS is a tool that can significantly reduce the delivered dose by choosing the most appropriate tube voltage while maintaining image quality in unenhanced head CT scans of infants.