This article presents an air pollution modelling approach and its use in health applications within the EO2HEAVEN project. The model is based on a multidimensional Inverse Distance Weighting and makes use of artificial distances between area attributes. The number and type of attributes used is fully customizable and can be adapted according to specific application fields and data preconditions. It is kept flexible and simple and thus, suitable to be used within a Spatial Data Infrastructure to provide access to realtime air pollution information via the internet. In a prototypical implementation the model is applied to estimate the concentration of particular matter and ozone in the Federal State of Saxony, Germany.