The Sleep Apnea Syndrome is known to be related with a partial (hypopnea) or total (apnea) collapse of the upper airway. The patency of the upper airway is dictated by a combination of passive mechanical properties and active neural mechanisms. From a physical point of view, neglecting active neural control mechanisms, the airway collapse can be explained by the interaction of the airflow during respiration (fluid) with the surrounding tissues (structure). The characteristic recurrent collapse causes about 3 % of the adult population to suffer from Obstructive Sleep Apnea (OSA). A systematic review of published studies on OSA insists on the necessity of a better understanding of the factors determining the airway obstruction in order to improve modelling, simulation and eventually prediction of the fluid-structure interaction at the origin of OSA. Prediction of OSA is in particular important to foresee the outcome of surgical procedures and to estimate its long-term effectiveness. In order to be of practical use a surgical prediction tool demands reliable and accurate simulation results available within a reasonable time delay and exploiting relevant and ‘in-vivo' accessible physiological data. In order to contribute to the understanding of OSA a physical approach is presented consisting mainly of three stages. At first, relevant physical quantities are formulated based on physiological ‘in-vivo' observations in order to define the problem. Secondly, a simplified theoretical model is proposed based on assumptions derived of the order of magnitudes derived during the first stage. Thirdly, an experimental validation of the assumptions and of the accuracy of the theoretical model is performed in a suitable experimental setup allowing to control model input parameters and to measure model output parameters.