Although repeated sprint tests are relatively common, there have been few investigations of repeated sprint exercise using non-motorised treadmill ergometry. The purpose of this study was to determine the reliability of a repeated sprint procedure using this apparatus. Ten healthy, active males, performed three repeated sprint tests (six repetitions of 6 s sprints with 30 s recovery) on three separate occasions. Performance as determined by maximal speed, average force production, and fatigue were compared across the three trials. Maximal speed and average force were not significantly different between visits (p < 0.05) and a variety of reliability measures suggested good agreement (e.g., coefficient of variations no more than 5 %). The fatigue indices for maximal speed and for average force were generally less reliable (coefficients of variation around 30 % in both cases). In conclusion, measures of performance (maximal speed and average force) can provide reliable results in a repeated sprint protocol but the reliability of fatigue measures appears to be low.