Objective. Exercise-heat strain estimation approaches often involve combinations of body core temperature (Tcore), skin temperature (Tsk) and heart rate (HR). A successful existing measure is the ‘Physiological Strain Index’ (PSI), which combines HR and Tcore values to estimate strain. However, depending on variables such as aerobic fitness and clothing, the equation’s ‘maximal/critical’ Tcore must be changed to accurately represent the strain, in part because high Tsk (small Tcore−Tsk) can increase cardiovascular strain and thereby negatively affect performance. Here, an ‘adaptive PSI’ (aPSI) is presented where the original PSI Tcorecritical value is ‘adapted’ dynamically by the delta between Tcore and Tsk. Approach. PSI and aPSI were computed for athletes (ELITE, N = 11 male and 8 female, 8 km time-trial) and soldiers in fully encapsulating personal protective equipment (PPE, N = 8 male, 2 km approach-march). While these were dissimilar events, it was anticipated given that the clothing and work rates would elicit similar very-high exercise-heat strain values. Main results. Mean end HR values were similar (∼180 beats min−1) with higher Tcore = 40.1 ± 0.4 °C for ELITE versus PPE 38.4 ± 0.6 °C (P < 0.05). PSI end values were different between groups (P < 0.01) and appeared ‘too-high’ for ELITE (11.4 ± 0.8) and ‘too-low’ for PPE (7.6 ± 2.0). However, aPSI values were not different (9.9 ± 1.4 versus 9.0 ± 2.5 versus; p > 0.05) indicating a ‘very high’ level of exercise-heat strain for both conditions. Significance. A simple adaptation of the PSI equation, which accounts for differences in Tcore-to-Tsk gradients, provides a physiological approach to dynamically adapt PSI to provide a more accurate index of exercise-heat strain under very different working conditions.