Until present visibility has been assessed as a Boolean phenomenon in various kinds of Geographical Information Systems (GIS) and Computer Aided Design (CAD) systems: either things can or cannot be seen. Probabilistic Visibility (PV) is presented as an advanced alternative. PV represents the probability that one location of object can be seen from another. Results of PV analysis are stored in Probabilistic Visibility Graphs (PVG) which stores the probability of visual contact for all relevant pairs of locations (eg raster cells). A visibility graph will be calculated based on field experiments and literature visibility decay-as a function of terrain, distance, viewing-angle and vegetation type-. The presentation takes its point of departure from analysis of visibility in forested environments. It describes the field experiment, and how the graph is calculated and visualized. Furthermore, perspectives of the further development and application of probabilistic visibility graphs will be given.