Background Demand is increasing for youth-specific preference-based health-related quality-of-life measures for inclusion in evaluations of healthcare interventions for children and adolescents. The EQ-5D-Youth (EQ-5D-Y) has the potential to become such a preference-based measure. Objective This study applied the recently published EQ-5D-Y valuation protocol to develop a German EQ-5D-Y value set and explored the differences between values given to youth health by parents and non-parents. Methods To elicit EQ-5D-Y health state preferences, a representative sample of 1030 adults of the general population completed a discrete choice experiment (DCE) online survey, and 215 adults participated in face-to-face interviews applying composite time trade-off (cTTO). Respondents were asked to consider a 10-year-old child living in the health states. DCE data were modelled using a mixed logit model. To derive the value set, DCE latent scale values were anchored onto adjusted mean cTTO values using a linear mapping approach. Results Adult respondents considered pain/discomfort and feeling worried/sad/unhappy as the two most important dimensions in terms of youth health. Adjusted mean cTTO values ranged from − 0.350 for health state 33333 to 0.970 for health state 21111. The EQ-5D-Y value set showed a logical order for all parameter estimates, and predicted values ranged from − 0.283 to 1. Differences in preferences by parental status were mainly observed for cTTO results, where mean values were larger for parents than for non-parents. Conclusions Applying the valuation protocol, a German EQ-5D-Y value set with internally consistent coefficients was developed. This enables the instrument to be used in economic evaluations of paediatric healthcare interventions.