Wheat yield and water productivity are affected by water and nitrogen fertilizer management and climatic data. The aim of the present study was to assess the efficiency of artificial neural networks (ANNs) models in predicting winter wheat yields and water productivity using climatic data and seasonal water-nitrogen variables. The calibration and verification of the proposed models were accomplished using data from field experiments carried out in three years. The models were evaluated using several statistical error and goodness-of-fit measures, including the coefficient of determination, root mean square error (RMSE), mean bias error (MBA), and standard error (STE). The applied ANNs models had desirable accuracy in estimating crop yields and water productivity. ANN models consistently produced more accurate predictions than multiple linear regression models. The sensitivity analysis indicated that the productivity of wheat water has the highest level of sensitivity to the seasonal water use. The data generated here suggested that maximum and minimum wheat water productivity would be achieved with about 272 and 490 mm seasonal water use, which were determined 1.55 and 0.81 kg m-3, respectively. The applied ANNs can provide the background for promoting the water productivity for this strategic crop in different weather condition, and give us the possibility of the logical and economical use of water sources and nitrogen and also programming the combined usage of both resources in the study area.