This study assessed the longer-term (12-month) variability in post-exercise heart rate recovery following a submaximal exercise test. Longitudinal data was analysed for 97 healthy middle-aged adults (74 male, 23 female) from 2 occasions, 12 months apart. Participants were retrospectively selected if they had stable physical activity habits, submaximal treadmill fitness and anthropometric measurements between the 2 assessment visits. A submaximal Bruce treadmill test was performed to at least 85% age-predicted maximum heart rate. Absolute heart rate and Δ heart rate recovery (change from peak exercise heart rate) were recorded for 1 and 2 min post-exercise in an immediate supine position. Heart rate recovery at both time-points was shown to be reliable with intra-class correlation coefficient values≥0.714. Absolute heart rate 1-min post-exercise showed the strongest agreement between repeat tests (r=0.867, P<0.001). Lower coefficient of variation (≤ 10.2%) and narrower limits of agreement were found for actual heart rate values rather than Δ heart rate recovery, and for 1-min rather than 2-min post-exercise recovery time points. Log-transformed values generated better variability with acceptable coefficient of variation for all measures (2.2–10%). Overall, 1 min post-exercise heart rate recovery data had least variability over the 12-month period in apparently healthy middle-aged adults.