Observed individual variability in cardiac baroreflex sensitivity (cBRS) and heart rate variability (HRV) is extensive, especially during exposure to stressors such as heat. A large part of the observed variation may be related to the reliability (consistency) of the measurement. We therefore examined the test–retest reliability of cBRS and HRV measurements on three separate occasions in 14 young men (age: 24 (SD 5) years), at rest and during whole‐body heating (water‐perfused suit) to raise and clamp oesophageal temperature 0.6°C, 1.2°C and 1.8°C above baseline. Beat‐to‐beat measurements of RR interval and systolic blood pressure (BP) were obtained for deriving HRV (from RR), and cBRS calculated via (i) the spontaneous method, α coefficients and transfer function analysis at each level of heat strain, and (ii) during forced oscillations via squat–stand manoeuvres (0.1 Hz) before and after heating. Absolute values and changes in all cBRS estimates were variable but generally consistent with reductions in parasympathetic activity. cBRS estimates demonstrated poor absolute reliability (coefficient of variation ≥25%), but relative reliability (intraclass correlation coefficient; ICC) of some frequency estimates was acceptable (ICC ≥0.70) during low‐heat strain (ICC: 0.56–0.74). After heating, forced oscillations in BP demonstrated more favourable responses than spontaneous oscillations (better reliability, lower minimum detectable change). Absolute reliability of HRV estimates were poor, but relative reliability estimates were often acceptable (≥0.70). Our findings illustrate how measurement consistency of cardiac autonomic modulation estimates are altered during heat stress, and we demonstrate the possible implications on research design and data interpretation.