Due to the scarcity of water resources worldwide, it is essential to determine the water productivity indices. In this study, the SWAP model was used to determine the agricultural water productivity indices for three wheat farms in the arid regions of Iran. The model was calibrated and validated for each study farm using a large number of field-measured data. The results showed that the model could satisfactorily predict moisture profiles. SWAP model calculated water productivity. Due to the results of this study, it is possible to increase wheat yield by 14%, if the irrigation scheduling is correctly planed. Deficit irrigation by 30% showed no significant effect on reducing yield. Appropriate irrigation scheduling has increased WPETdp (yield to crop actual evapotranspiration plus deep percolation ratio) and WPIrr (yield to total applied water ratio) by 48 and 61%, respectively. High evaporation at the initial stages of growth decreases WPET (yield to crop actual evapotranspiration ratio) by 28% compared to WPT (yield to crop actual transpiration ratio). Improving agricultural operations such as mulch or soil application using subsurface irrigation methods can improve WPET. Reducing the applied irrigation depth had a negligible impact on the WPET and WPT indices, but the WPETdp and WPIrr indices exhibited a significant increase.