Today, the water demand and surface water shortage has increasing, conjunctive use is known as an approach to overcome on problems. Toward this, using simulation/optimization models is a powerful technique to optimize the Operation of surface and groundwater systems. Simulation models are able to simulate system responses to applied management policies. Optimization models, determine the best management policies among the set of possible management policies. In this research, artificial neural network as a simulation model and Ant algorithm as an optimization model are used to determine optimal operating policies in Najafabad Plain (Zayandeh-Rud Sub Basin). Knowing of existence of three modern irrigation network in the study area, the management model is run for three agricultural area; Nekouh Abad Rast, Nekouh Abad Chap and Khamiran. 240 rows of data from the 20 years data are used for training artificial neural networks. The ANN models concluded that there are able to estimate water table levels in agricultural areas, under monthly water management policy. Then, by combining these models with Ant optimization algorithm model, Optimum water policies for reducing water shortages are obtained. the results, shows water resources shortage for agricultural needs in the study area, so water demand should reduce to achieve sustainable development. Also, Determination the allocation of surface water and groundwater in drought and wet periods for each area is the results of this thesis. Keywords: Water resource management, Conjunctive use, Simulation, Artificial neural network (ANN), Optimization, Ant system, Najafabad plain , Groundwater