The purposes of this study were (I) to assess the reproducibility of endurance performance testing on an air-braked cycle ergometer, and (II) to compare laboratory performances to performances in road races. Ten well-trained, competitive cyclists (peak power output [PPO] 443 +/- 37 W, [values are mean +/- SD]) undertook either: (I) three 20 km and three 40 km time trials (TT) on an air braked ergometry system (Kingcycle) (n = 6), and/or (II) three 40 km laboratory TT and two 40 km road TT competitions (n = 8). The time taken for the laboratory simulated 20 km and 40 km TT rides were highly reproducible (coefficient of variation 1.1 +/- 0.9% and 1.0 +/- 0.5%, respectively). However, the mean power output and heart rate were significantly different (p < 0.0001) between the 20 km and 40 km TT (327.5 +/- 16.9 vs 303.9 +/- 14.9 W and 171.4 +/- 5.1 vs 168.3 +/- 4.4 beats/min, respectively). A strong relationship (r = 0.99, p < 0.001) was observed between the mean cycling time and the average sustained power output. A significant correlation (r = 0.98, p < 0.001) was also observed between laboratory and road race times, although road race times were, on average, some 8% slower. These findings indicate that the Kingcycle ergometry system can be used as a reliable method of assessing short term endurance cycling performance.