Nowadays due to population growth, climate change impacts, hydrological uncertainty and increasing water requirements, more than ever, society needs to have accurate integrated management to supply its demands in different parts such as: agriculture, hygienists and industry. One of the fundamental steps to have an integrated water resources management in a basin wide and supplying its demands is the optimum conjunctive use of surface water and groundwater. In this regard, in this study, optimum utilization of surface water and groundwater resources is applied in Najafabad plain in Gavkhooni basin which is one of the most important sub-basin because of these reasons: (1) its contribution to supply the agricultural needs of the basin, (2) having a negative balanced problems and loss of quality and quantity of groundwater resources, (3) having certain complexity in terms of nutritional conditions and interaction between surface water such as Zayandehrud River. To solve the problem, simulation-optimization method using Artificial Neural Network model for simulating and Honey- Bee Mating algorithm as optimization model, was applied. After training Artificial Neural Network model with 276 rows of data from 23 last years, optimization model was developed due to different constraints such as: water resources capacity, drawdown of water table in aquifer, maximum amount of surface water and ground water. To create optimal utilization model and surveying different operational policies, after linking simulation model and optimization models, 2 separate operating policies which include 3 scenarios with different climatic conditions were developed. Finally, according to the allocation values which obtained from the first operating policy, we tried to modify crop cultivation dominant pattern to maximize net income by using GAMS software, under different constraints such as: minimum and maximum arable area for each crop, maximum total cultivable area in each region and available amount of surface water and groundwater. The results show that selected simulation model with R 2 above 95 percent and less than 8 percent error in the validation of the model for forecasting has a good performance to simulate the aquifer behavior. Also results of the first scenario show that the model can't supply all the needs due to existing constraints. However it can improve the mean groundwater level to the acceptable. In the second scenario the model would able to improve the mean groundwater level with temporal distribution of available water resources with getting help from the excess surface water at normal and wet years in the both agricultural zones. Keywords: Water resource management, Surface water, Groundwater, Simulation, Artificial Neural Networks, Optimization, Honey Bee Mating Algorithm, Optimum cropping pattern.