This study examined the stability of children's cortisol responses to a social evaluative laboratory stressor from early to middle childhood. Ninety-six children (51 males) completed stress-inducing laboratory tasks and provided five salivary cortisol samples in early (W1) and middle (W2) childhood. Although W1 cortisol responses did not predict W2 cortisol responses, children's cortisol responses demonstrated change: compared to their W1 cortisol responses, children's W2 cortisol responses demonstrated an increased slope and more negative quadratic curvature. Furthermore, child psychiatric symptoms at W1 moderated the stability of children's cortisol responses. Children with fewer preschool psychiatric symptoms demonstrated greater inter-individual and intra-individual stability, whereas children with higher preschool psychiatric symptoms and comorbidity demonstrated systematic inter-individual and intra-individual instability in cortisol responses over time. Findings suggest a developmental shift toward increasing cortisol stress responses from early to middle childhood and highlight preschool psychopathology as a moderator of stability in children's cortisol responses over time.