Background: Psychophysiological hyperresponsiveness to social-evaluative stress plays a key role in current theories of social phobia (SP). Owing to the early onset of this disorder, the study of children with SP can help to improve etiological models. However, research to date has failed to clarify whether children with SP are physiologically hyperresponsive to stress or not. In addition, the existence of elevated pre-stressor baseline group differences complicates the interpretation of acute stress responses and therefore poses a challenge for this line of research. Drawing on current models of autonomic control to explain the relationship between baseline and stress responding in SP children and healthy controls (HC), we acquired a broad set of autonomic measures in our study. Method: To index baseline and stress reactivity and to assess autonomic flexibility, we assessed a comprehensive array of sympathetic and parasympathetic measures in SP children (aged 8 to 12 years; n = 30) and healthy control children (HC, n = 26), while exposing them to the Trier Social Stress Test for Children (TSST-C). Results: At baseline, the SP children showed higher levels of sympathetic (heart rate, electrodermal activity) and lower levels of parasympathetic (respiratory sinus arrhythmia, RSA) activation when compared to the HC children. During the TSST-C, the SP children showed similar HR responses but a limited RSA reactivity and a slower HR recovery relative to the HC children. Conclusion: Our study extends previous research by showing elevated baseline arousal and comparable stress responding in SP children relative to HC children. In addition, based on the autonomic flexibility model, we provide a potential explanation for the null findings of previous studies during stress. The pattern of elevated baseline heart rates and reduced RSA point to restricted autonomic flexibility in children with SP.