Literature on restrained eaters posits that the capacity to successfully diet is cognitively demanding, requiring exertion of cognitive control over one’s eating behaviour. Demands that deplete this limited inner resource can thus lead restrained eaters to become disinhibited and consume large amounts of food. Although self-regulation is inferred to play a role, it has not yet been studied in combination with its physiological correlates. As low heart rate variability (HRV) is associated with lower self-regulatory capacity, and acute mental stressors decrease HRV, it was hypothesized that reduced HRV during stress may mediate the relationship between stress and food intake, and that the mediation may be moderated by dietary restraint. Specifically, the moderation hypothesis predicted more restrained eaters would exhibit a stronger relationship between low HRV and intake. Female undergraduates ( n = 92) were randomized to a stress or control condition while HRV was recorded. Participants subsequently engaged in a bogus taste test to quantify post-stress consumption. Restraint was conceptualized using Hagan and colleagues’ (2017) latent restraint factors. A significant moderated mediation emerged for Weight-Focused Restraint, ab3 = .196, SE = .077, 95% CI [.060, .361] and Calorie Counting, ab3 = .062, SE = .023, 95% CI [.024, .115]. However, contrary to expectation, stress-induced HRV reduction was associated with decreased intake for those low in restraint, whereas intake was unrelated to HRV for individuals with higher dietary restraint. Further, there was no evidence of traditionally-defined disinhibition in more restrained eaters. Results suggest those with higher restraint show a relative disconnection from internal physiological cues shown to affect intake under stress in less restrained individuals.