The objective of this study was to evaluate the role of ASTER and digital elevation model (DEM) data for classifying soils and agricultural lands. The obtained results from the image processing and spectral reflection revealed that the green plants have higher reflection in near infrared than wilted and bare soils. The results also demonstrated that that the highest reflections in visible and middle infrared are mainly related to the saline and gypsiferous soils. Besides, the lowest reflection of the saline soils is attributed to for the semi-wet soils with bright salt crusts. It is due to the moisture effect and sponge like surfaces of the crusts that can be due to a great absorbtion in sun radiations. All ASTER bands are able to separate the stony, rough and uneven land classes from the homogenous saline soils wit the soft and black appearance. Based on the obtained results we may generally conclude that for separating the saline classes with different surface characteristics digital elevation model (DEM) has a key role in improving accuracy. This result also indicated that the visible and near infrared bands can be mainly used for separating the salt crust soils from some other soil classes.