Maximal oxygen uptake (VO2 max) is difficult to measure and most predictions are inaccurate due to a variety of assumptions. The purpose of this study was to validate a dynamical system model (DSM) for predicting HR max and VO2 max during walking and running. A secondary purpose was to predict VO2 responses using a neural network. Twenty-six healthy males completed a maximal cardiopulmonary exercise test (CPET) and a submaximal protocol. The models were applied to the submaximal data to estimate the participants’ HR/VO2 responses and predict their HR max and VO2 max. The model accurately tracked HR and VO2 responses (R2 = -.85-0.99). However, it did not accurately estimate max (R2 < 0). Further refinement of the model is needed. This study elucidated some of the challenges of using a DSM and demonstrated that a neural network may be useful for easily predicting VO2 responses.