Geostatistical methods were successfully used in detailed modeling of the two most important and significantly different Upper Pannonian and Lower Pontian reservoirs in the Kloštar Field. The research was aimed at construction of a detailed geological model of these two sandstones, which should represent typical model (in terms of lithology, granulometry and depositional conditions) for the analysis of the Upper Miocene sandstone reservoirs in the Sava Depression. The combination of the three variables (porosity, thickness and depth of the reservoirs) has helped to map the position of a depositional channel, different lithofacies, detritus transport direction, local palaeorelief influence, variables distribution within the reservoirs and mapping areas of greatest uncertainty. The reservoirs have been analyzed with deterministical interpolation methods (Ordinary and Indicator Kriging techniques) and stochastical methods (Sequential Gaussian and Sequential Indicator Simulations). Kloštar anticline was mapped using Ordinary Kriging. The Sequential Gaussian Simulations have given excellent results in defining variables distribution as well as in locating the areas of the greatest uncertainties. It was proven that during the Late Pannonian the main material transport direction from the Eastern Alps by turbidites was northwest-southeast, whereas during the Early Pontian it has changed to north-south direction. Those directions have been confirmed locally in the Kloštar area by the probability maps obtained by the Indicator Kriging technique and the Sequential Indicator Simulations for the porosity and thickness variables. The maps clearly show distribution of various lithofacies ranging from pure sandstones to basin marls, and also indicate the material transport direction. However the question of the influence of Moslavačka Gora Mt. as a local material source still remained unsolved.