The reliability of heart rate responses to non-steady-state tasks among 37 men with spinal cord injuries (lesion level: C4/5-L5) was examined with a simple heart rate recording device (Sport Tester PE3000). Three identical trials of 6 different transfers and an 8-cm curb ascent were performed on one day (Trial 1 and 2; n = 37) and one week later (Trial 3; n = 12). Pearson's r and intraclass correlations for the highest and the mean heart rate provoked during Trial 1 and 2 ranged from 0.73 to 0.97 for the transfers and from 0.92 to 0.97 for the curb ascent. Correlations were somewhat lower for Trial 1 versus Trial 3. A paired t-test revealed lower heart rate responses to Trial 2 and 3, suggesting a moderate learning effect and/or a reduction in psychological stress. It was concluded that heart rate responses to non-steady-state tasks, as recorded by a Sport Tester PE3000, are reproducible in men with spinal cord injuries.