||Good nutrition habits are a viable health aspect. The goal of this project is to continuously monitor human nutrition phases and detect malnutrition from different body-worn, unobtrusive and non-invasive sensors. We present the investigated on-body sensing domains and their discrimination properties. By fusing information from independent sensing domains a context recognition system is envisioned that can automatically monitor nutrition related details. The expected information derived from this system include eating habits, e.g. rate of intake, duration and daily schedule as well as estimations of nutritional value, all of which can be relevant health issues. We expect that this system in combination personalised coaching from health professionals will help significantly to avoid different forms of malnutrition and achieve / maintain a preventive lifestyle.