The Trier Social Stress Test (TSST), the gold standard for inducing psychosocial stress in the laboratory, has been successfully transferred to a virtual setting in recent years. Whether the transferability using virtual reality (VR) also applies to a child version, the TSST-C, has not yet been evaluated. The aim of this study was to compare the multimodal stress response of a real and a virtual TSST-C. For this purpose, N = 73 healthy children and adolescents aged between 11 and 17 years were assessed. One group (n = 37) underwent a TSST-C in a real environment, the other group (n = 36) a TSST-C in a virtual environment. Salivary cortisol was collected as an endocrinological stress parameter, salivary alpha-amylase, heart rate, and heart rate variability as autonomic stress parameters, and subjective responses as psychological stress parameters. The experience in VR (presence) and possible negative effects (VR sickness) that could bias the stress response were also recorded. In addition, various parameters for the demographic and psychometric comparability of the groups were recorded and controlled. Significant differences between the groups in cortisol and subjective stress were identified, but not in the parameters of the autonomic nervous system. There was equivalence between the groups for alpha amylase. The participants’ degree of presence had no significant influence on the experience of stress. The benefits and limitations of a virtual TSST-C are discussed.