Abstarct: Identifying hydrocarbon-producing zones and separation of the boundary between the oil and water to the perforation and exploitation of oil wells, is one of the key steps in petroleum engineering. For this purpose usually applied cut-off methods on petrophysical parameters such as porosity, saturation and shale volume is used, which has wrong.In addition, one of the most important problems in engineering, uncertainty and the confidence to respond.In this thesis a new Approach of producing hydrocarbon zones Sarvak Formation in the southwestern Iranian oil fields have been studied and identified. For this purpose, in the first stage of this formation reservoir properties including porosity, permeability, saturation and shale volume as the main factors controlling hydrocarbon production in this field was determined.In the next step, according to the conventional method and applied cut off on the petrophysical parameters, pay zones were identified.Then, using the conditional Bayesian probability theory and the theory of evidence as the theories presented on data fusion evaluated and these zones were determined.Compare the results with the results of well testing has shown that relying on a data fusion, identify pay zones with higher precision is carried out.also uses the theory of evidence in addition to achieving this purpose, by specifying the uncertainty in the results, the possibility of more precision reservoir modeling is provided and confidence to the results will be investigated quantitatively.