Most existing physiological models were designed for either very short periods of intense exercise, i.e. sprinting, or for long periods of constant exercise, such as cycling time-trialing. However, some competitions, for example the cycling team pursuit, require athletes to repeatedly modulate their exercise intensity between high and medium level power outputs during a short event. This study takes an experimental approach to investigate the applicability of existing physiological models to intermittent exhaustive exercise of short duration (approximately 5 minutes in length). Several physiological models were identified as potentially suitable and then 6 national level cyclists carried out a series of tests to determine which of these models was most applicable to short intermittent periods of exercise. The tests consisted of 3 stages. The first stage consisted of 3 different constant power exhaustion tests, from which it was possible to characterize the athlete’s power-duration curve. In the second testing stage the athletes took part in 4 micro-interval exhaustive tests that were focused on investigating the recovery and fatigue of the athletes. The final stage of tests were mock ‘Team Pursuit’ tests that were designed to assess the validity of the different models. All the tests took place on a Lode Excalibur Sport PFM cycling ergometer. The results suggest that some of the models cannot be applied to intermittent exercise and that in order to accurately model short intermittent exhaustive exercise recovery and fatigue must be accounted for.