Theorists argue that parental reflective functioning (PRF) is activated in response to emotions, potentially supporting parenting sensitivity even when arousal is high. That is, when parents become emotionally reactive when interacting with their children, those who can use PRF to understand their children’s mental states should be able to parent sensitively, which, in turn, should promote children’s ability to understand their own mental states. We test this theory by examining whether, in the face of physiological reactivity, mothers’ PRF inhibits one form of parenting insensitivity, overcontrol (OC), and whether this process in turn predicts children’s RF. A diverse sample of school-age children (N = 106, Mage = 10.27 years) completed a standardized failure paradigm while their mothers were asked to passively observe. Following the stressor, mothers and children independently completed interviews regarding the task, which were later coded for RF with respect to children’s mental states. Mothers provided saliva samples before and after the stressor, and after the interview, which were later assayed for cortisol reactivity; maternal behavior during the stressor task was coded for OC. Among mothers with low levels of RF, greater increases in cortisol were associated with more displays of OC, whereas among mothers with high PRF, greater cortisol reactivity was associated with fewer OC behaviors. For low PRF mothers, higher reactivity and OC predicted lower children’s PRF for their own experiences. The findings provide initial evidence for a protective function of PRF, and may point toward the importance of promoting PRF in intervention programs to reduce parental OC.