Geothermal energy is attractive as an alternative base load supply due to its virtual inexhaustibility, its low emission of carbon dioxide, and its continuous availability. However, the discovery risk is one of the main barriers to the use of hydrothermal geothermal energy in Germany. Minimising this risk is the aim of the presented research project. In a first step, the project compiles a statistically significant data base of thermal and hydraulic properties of subsurface rocks in Germany in order to facilitate and improve the design of geothermal installations in the planning stage. Data from direct measurements on core samples are combined with the interpretation of geophysical borehole measurements to derive statistical moments of the desired properties for selected main rock types as well as for geologic strata in the German subsurface. In a second step, the uncertainties of rock properties at a given location and for a target parameter, such as temperature or flow rate, are quantified. This stochastic approach aims to obtain not only average values and error estimates of the target parameter, but also its spatial distribution. Based on this information, the risk within a geothermal project can be estimated more precisely. As a result, costs are reduced or estimated with less uncertainty. As an example of this method, an exploration scenario is presented for a geothermal heating project in The Hague, Netherlands, where the predicted uncertainty in reservoir temperature could be reduced by more than 50 %.