The efficiency of the selected regression model to predict and map the spatial variation in soil salinity is shown by the strong relationship (R2 = 0.78) at the 95% probability level, and RMSE of 0.54 dS/m. This is partially due to the result of the selected remote sensing soil salinity indicators. The selected model in this study showed superiority in the prediction power of soil salinity over those reported by Shrestha (2006) (R2 = 0.23). Moreover, the good performance of the selected model in this study was due to the enhanced image efficiency in highlighting information from soil salinity and suppressing other details.