A model is a representation of a real system to describe some properties i.e. internal factors of that system (out-puts) as function of some external factors (inputs). It is impossible to describe the relation between all internal factors (if even all internal factors could be defined) and all external factors (if also even all these factors could be defined). Therefore, before building a model, a decision has to be made about both the properties of the system that have to be known for the problem under concern and the relevant external factors. In general, the model will be a mathematical expression and will range, dependent on the application, from relatively simple single input-output models to more complicated multi input-output models. The result can be a black box model, based on curve fitting of experimental data of considered factors. Such models are interesting if only few factors have to be considered and they can be very usefull for comparative studies or for the calculation of changes of a factor instead of the absolute level of it (e.g. in optimization studies). One has to be aware that these models are only valid within the range of the experimental conditions. If one tries to understand and analyse the relation between the internal and external factors then a physical approach can be added. In this approach causal relationships are formulated between the various factors. This visualizes the link between various factors and it is a powerfull tool to analyse the system. Because one has to restrict in the amount of internal and external factors, the physical-mathematical model still is a projection of the real system. Application of this kind of models beyond the range of experimental verification seems reasonable but one has to be conspicious. In greenhouse engineering both black box and physical models are developed, as will be discussed.