Fracture intensity indicators from well data are highly uncertain when applied or extrapolated to the whole field. Statistical techniques offer some useful tools for dealing with such problems in reservoir modeling. The recent research and development of effort focuses strongly on the development of hybrid techniques that aim at overcoming the inherent assumptions and limitations of individual techniques and taking into consideration uncertain data for estimating error bars (or confidence bounds) on the calculated outputs. This paper presents a recently developed hybrid technique based on soft computing methods and shows how this technique can be used to improve the workflow of fracture characterization and sweet-spots identification. It uses a set of rock matrix properties and a fracture-related seismic attribute to simulate a three-dimensional fracture intensity model at the Pinedale anticline in Wyoming. The estimated fracture intensity results agree very well with the data at four virtual wells. It also shows that use of the fracture intensity estimations, together with their error bars, is a valuable tool for the identification of potential sweet spots with reduced risks.