This work analyzes the feasibility of building a three-dimensional fluvial model by combining geostatistics and geological processes simulation. This methodology was applied for clastic reservoirs prognosis and estimation in the Mina del Carmen Formation in the El Trebol oilfield, Golfo San Jorge basin. As a first step, a geologic process modeling technique was used for creating a valid numerical model which honored the conceptual model. Different descriptive parameters of geological processes were set to control the simulation of sand and reservoir deposition. The results were qualitatively compared with a seismic cube, and quantitatively validated with sand fraction, areal and vertical trends. As a second step, a geostatistical approach was used to match spatial sand distribution with well data. Sequential Indicator and Multiple Point Statistics simulation methods were used with the process simulation model as a trend. Both results were compared, and the whole process was validated with blind tests over three wells. The Multiple point simulation showed better sand prediction and spatial continuity. This case study showed that the combined two methodologies turned out with acceptable results and achieved benefits from both. In this way, a good match to well data and trends was obtained due to the geostatistical methods, and besides geoforms and conceptual model were better captured due to the process simulation. This combined methodology can improve forecasts, specially at initial stages in oilfield development, with a low number of wells and big uncertainty.