Abstract: Evaporation is one of important parameter to determine crop water requirement especially in arid and semi-arid regions, therefore the accurate calculation of this factor is employed for research's such as water balance, irrigation system design, and water resource management is very important for these applications, In order to calculate the daily evaporation values, direct measurement or neural network models can be employed for the regions which there is no direct measurements tools. In this paper the daily evaporation has been estimated by using daily temperature, relative humidity, wind velocity, sunshine hours, and evaporation data in Abu-garib Automatic weather station employing Neural Network training using daily data used four years and network testing takes one year. In this approach , a MLP layer network with a hidden layer and sigmoid function has been employed. The results show the powerful capability and high accuracy of artificial neural networks in calculating of daily evaporation. Best model for estimation of evaporation is neural network have MSE (0.1515), AE (0.2892) and R (0.9909).