The aim of this thesis was to develop a mathematical model able to simulate heart rate during cycling exercise with power as input. From an optimization of free model parameters to fit the model to heart rate recordings, we wanted to extract information that possibly could indicate the health of subjects. Our model is based on physiological principles, and three physiologically interpreted parameters are being optimized. The parameters τL and νmin are assumed to predict the relative condition of a subjects, while the last parameter p ̄ T is as- sumed to predict the absolute performance of a subject. Maximal heart rate is the only subject dependent parameter that has to be known a priori as input to the model. Results in this thesis indicate that the model predicts parame- ters consistent within subjects, and parameter variations are observed between subjects. However, a further validation of the model is needed to decide if all parameters change as expected according to health of subjects, such that it can be used as a tool in training analysis.