Abstarct: Providing a precise fracture model and distributing them in naturally fractured reservoir is one of the most complex cases in the engineering of reservoirs. Fracture modeling is only baxsed on well data with a low impact range and hence leads to high uncertainty. Conventional fracture modeling techniques usually use intrusion detection algorithms to predict the spatial distribution of fractures. In this research, using fracture controlling factors, the fracture distribution model is obtained by combining three-dimensional seismic data and wells. For this purpose, after determining the fracture zones and their orientation using well data, the fracture determinant attributes were extracted from seismic data. These attributes have been used as fracture controller to build fracture distribution model. Fracturing algorithms were implemented on seismic data and fracture distribution models were co-motivated by collocated co-kriging method on attributes. Then, by classifying the fractures of the reservoir, a discrete fracture network model was obtained for different scales. The comparison of the discrete fracture network obtained with the proposed strategy with the conventional method showed that fracture controlling factors as fracture stimuli for distributing fractures in the reservoir can be obtained by using 3D seismic data in There are cases where there are fewer wells in the reservoir. The proposed strategy was implemented on a fractured reservoir in one of the Gulf fields, and a discrete fracture network was obtained more accurately than the conventional method. But due to the complex nature of fractures and the lack of data associated with them, the resulting models are always associated with uncertainty. In the second phase of this research, a new method for constructing a fracture model is used to determine the model made in the first stage. This modeling is done through the relationship that may exist between fracture intensity and fracture drivers. Artificial intelligence tools are used to establish a correlation between fracture intensity and a group of fracture drivers. Fractures controlling factors such as porosity, lithology, distance from faults and seismic attributes including relative acoustic impedance, ant tracking algorithm, edge enhancement, chaos, curvature, and variance were considered as fracture drivers. We used the neural network to establish a relationship between fracture intensity and fracture drivers. For modeling, the supervised method was used. Also, all fracture controlling factors were monitored with fracture density data in wells. For training the model, 100 and 500 iteration were used. The relationships were compared with two different replications and we found that with 500 iteration, a good correlation was found between the fracture intensity and fracture drivers, which was eventually used to construct a discrete fracture model. In the end, we came to the conclusion that in addition to the fact that the controlling factors of the fractures, with the absence of sufficient wells, play important and important role in the fracture modeling, a very good distribution of fracture intensity In the reservoir they offer.