In the last five years about 30 deep geothermal wells have been drilled in the Southern German Molasse basin. 16 of them have been planned and/or supervised by Erdwerk GmbH who supports the operator as a consultant throughout the project. One of the main duties of a consultant like Erdwerk is cost planning and time schedule forecasting for the well construction process. To this day, these estimates have been based on a historic average time for the main operations which have been added up to the total well construction time. Uncertainties have been taken into account by adding a contingency factor. This approach has the advantage of being simple, fast and easy to communicate. However, it does not give any idea about the variability of the estimate and the risks involved, which limits its application. Therefore, the aim of this master thesis was to establish a well construction model based on statistical methods to allow probabilistic time and cost estimation. Firstly, a review on probabilistic methods in well planning and revision of required statistical methods and concepts were performed. Then a model to determine the total well construction time was set up. Offset data of 16 wells was gathered and analyzed to determine a probability function for the duration of each process. The model was verified by comparison with real historic results. Trends, observed in the offset data, were implemented to model the performance mean and its variation over time. Then a multi-well model was established. Finally, the model was extended by adding costs. After the model set-up, a simulation run for a well with average casing setting depths was performed for each category of wells. The model architecture and the results are discussed critically. Underlying assumptions and simplifications are reported. The resulting constraints are described. A sensitivity analysis was performed to investigate the correlation strength of the different input distributions. Time vs. depth and cost vs. depth curves were drawn to visualize the simulation output. With the presented approach of well construction modeling, Erdwerk can deliver risk assessment for geothermal wells to investors, insurance companies and decision makers. This will aid proper budgeting and the calculation of insurance premiums. Moreover, the modeled technical limit or best historic performance can be used as technical performance reference. Based on the results of the sensitivity analysis, the key driving forces can be identified. Therefore, optimization strategies can be steered into the right direction.