A thorough knowledge of land use changes is important for planning and management activities of land resources.  Moreover, it is considered as an essential element for modeling and understanding of the major land forms, especially in arid regions like Saudi Arabia.  This study aims to assess the temporal changes in land use patterns due to the new developments (urban, industrial, commercial and agricultural) in Dirab region, Riyadh, Saudi Arabia using Landsat TM/ETM+ data and Erdas Imagine remote sensing software program.  Irrespective of source (active, passive and temporal), radiation detected by remote sensors travels a specific distance or path through the atmosphere which causes radiometric distortions.  To overcome this problem, images were radio-metrically (Top of Atmosphere) corrected applying a sun elevation correction and earth-sun distance techniques.  Radiative transfer equation was also used to compute the at-sensor radiance (Lrad) to surface radiance (i.e. absolute radiance value) applying pre-launch calibration constants of TM and ETM+.  Corrections for spectral emissivity were then made according to the nature of land use involved.  Supervised and unsupervised classification techniques were both applied to assess the land use class according to USGS level – I classification scheme using Erdas Imagine software.  Supervised classification was carried out using maximum likelihood method with the aid of ground truth data obtained from the field. However, un-supervised classification was made using algorithms that examined the unknown pixels in an image and aggregated them into a number of classes based on natural groupings or clusters presented in the image value.  Overall accuracy and error matrix with Kappa co-efficient are presented in the study, along with a change detection matrix of the study area.