Extracorporeal perfusion is the standard technique in cardiac surgery. It is controlled by perfusionists on the basis of their clinical experience and on the available data collected pre- and intra-operatively. But in spite of intensive monitoring postoperative complications occur. An appropriate control of the heart-lung machine (HLM) using an "autopilot" might improve the quality of heart-surgery and decrease postoperative complications. Hence, a mathematical model of a human circulatory system has been developed which provides much more information about haemodynamics, blood gases and acid-base status than standard monitoring. It has been implemented on a system which is capable of integrating measured data as input parameters in real-time in the simulation. Now, soft- and hardware control concepts based on the human circulatory system have to be developed which are able to control the HLM.