Abstarct: The term hydraulic fracturing refers to the process of initiating and expanding the fracture in the rock due to the hydraulic pressure applied by the fluid. The energy of conventional hydrocarbon reservoirs is over, while only about 30% of the oil is removed at the initial stage . Increasing the recovery and improving the recovery coefficient of the reservoirs by using hydraulic fracturing is one of the most effective ways for optimized productions. For better effectiveness of the hydraulic fracturing, chosing the type of well and the candidate laxyer is very important., Usually the purpose of selecting a candidate is to select single or group of wells or zones for the operations that are most likely to succeed. Most carbonate reservoirs have a low permeability, That’s why they can only produce economically if their hydraulic fracturing occurs. In this research, three wells in the Bangestan reservoir of Ahwaz oilfield were investigated for modeling the appropriate zones for hydraulic fracturing in Ilam and Sarvak Formations using a three-dimensional neural network method. In order to do this, we first developed a one-dimensional geomechanical model in these wells using petrophysical logs and internal tests. Having only a shear wave in one of the existing wells is the main problem in constructing a geomechanical model in these wells. In this research, one-variable regression was used to estimate this parameter in other wells. Thus an acceptable result of this method was extracted. After constructing the geomechanical model in the wells, by examining the obtained results and studying the previous researches on how to choose well and the candidate laxyer in the carbonate reservoirs of Iran, the criteria of the index of fragmentation and total fragmentation used to evaluate the laxyer of the data found in the Ilam and Sarvak Formations were used to make the 3D fracability model. The fragmentation index shows how much hydraulic fracturing can be effective on a particular laxyer. baxsed on existing data and previous studies in Iran's carbonate reservoirs, the relationship that was used to evaluate the fragmentation index in the Bangest reservoir of Ahwaz Square were used in this study. This index is a function of minimum horizontal stress, horizontal stress difference, Young modulus, fragility and internal friction angle that can be used in other carbonate reservoirs in Iran. Finally, 3D models of fracture ability were modeled using a multi-laxyered neural network method by using the criteria used by Sarvak and Ilam Formation and then suitable laxyers and wells for hydraulic fracturing function were identified and introduced. Results showed us an increase in the capacity of the reservoir from the south to the north as well as from the east to the west, and also this index decreases as the depth increases. Therefore, the most appropriate part for the hydraulic fracture is located in the upper and eastern laxyers of the Elam Formation.