A novel model for the coupling between ventricular repolarisation and heart rate (QT/RR) is presented. It is based upon a transfer function (TRF) formulism which describes the static as well as the dynamic properties of this coupling, i.e. the behavior after a sudden change of heart rate. Different TRF models have been analyzed by comparing their capability of describing experimental data collected with 19 healthy volunteers and several protocols of RR stimulation: i) rest with deep breathing at 0.1 Hz; ii) tilt with controlled breathing at 0.1 Hz and 0.33 Hz; and iii) cycling. The search for the best TRF leads to unambiguous identification of a three-parameter model as the most suitable descriptor of the QT/RR coupling. Compared with established static models (linear, or power-law), our model predictions are substantially closer to the experimental results, with errors 50% smaller. The shape of frequency and step responses of our transfer functions is essentially the same for all subjects and protocols. Moreover, each transfer function may be uniquely identified by three parameters, obtained from the step response, which are believed to be of physiological relevance: (i) gain for slow RR variability; (ii) gain for fast RR variability; and (iii) time during which QT attains 90% of its steady state value. The transfer function successfully describes the behavior of the RR control following an abrupt change of the RR interval, and its parameters may offer a tool for detecting pharmacologically induced changes, particularly those leading to increased arrhythmogenic risk. Key words —QT/RR coupling, transfer function