Abstarct: Natural fractures provides complex conduits for fluid flow, so it will affects on reservoir characterization, production rate and recovery factor. In this study, one of the southwest oil fields of iran which is naturally fractured carbonate reservoir getting evaluated in order to improve its fractures distribution. Fluid flow prediction of carbonate reservoirs is always challengable due to their complex and hetregeneous fracture network. Therefore, the results of fractures modeling can be used as input parameters for fluid flow simulation. Fractures network study is important to realize how to propagate and model fractures. The discrete fracture network is one of the fractures modeling methods. In this method, the continuous information can be import to the discrete model. In the first step, discrete fracture model will be generated within the train wells in reservoir by using fracture data obtained out of image log interpretation; but because of complex nature of fractures, the generated models are always associated with uncertainty. According to the elastic modules and geomechanical parameters evaluation, fracability index will asses the generation potential and propagation of fractures per depth. In the next step, by evaluation of petrophysical data and determination of fracability index through the reservoir area, re-zoning of fractures and conditiong them to the fracability are put into the agenda. In this way, the generated model will be more suitable with more certainty in detection of fractured zones through the wells which have no image logs. The overlapping depths of both primary model and the model conditioned baxsed on fracability in comparison with the generated model baxsed on image log interpretation which represents the real place of fractures, it was concluded that fracability baxsed model was able to detect the fractured zones with a 50 precent improvement in most of the zones.