Cortisol and heart rate variability (HRV) are well-established biomarkers of the human stress response system. While a relationship between cortisol and HRV is assumed, few studies have found evidence of their correlation within single study designs. One complication for isolating such a relationship may lie in individual variability in the cortisol response to stress such that atypical cortisol responding (i.e., elevated or blunted) occurs. To-date, studies on the cortisol response have employed traditional mean-difference-based approaches to examine average magnitude change in cortisol over time. Alternatively, data-driven trajectory modelling, such as latent growth mixture modelling, may be advantageous for quantifying cortisol based on patterns of response over time. Latent growth mixture modelling was used in N = 386 adults to identify subgroups based on trajectories of cortisol responses to stress. The relationship between cortisol and HRV was tested within subgroups. Results revealed a \'prototypical\' subgroup characterised by expected rise and fall in cortisol response to stress (n = 309), a \'decline\' subgroup (n = 28) that declined in cortisol after stress, and a \'rise\' subgroup (n = 49) that increased in cortisol after stress. Within the \'prototypical\' subgroup, greater HRV during stress was associated with decline in cortisol after stress from its maximum (r (306) = 0.19, p < 0.001). This relationship failed to emerge in the \'decline\' and \'rise\' subgroups (p > 0.271). Results document different patterns of cortisol response to stress; among those who exhibit a \'prototypical\' response, changes in HRV during stress are related to changes in cortisol after stress.