Land use and land cover (LULC) changes in many developing countries are inevitable. In Tabuk, a region northwest of Saudi Arabia, rapid urbanization and an agricultural evolution have led to dramatic changes in LULC over the last 30 years. Increasing land demand to meet food production desire has led to expanding these agricultural areas into previously undeveloped deserted or barren areas. This places immense pressure on valuable water reserves, therefore correctly defining irrigated or non-irrigated agricultural land is important to meet the need for intensive and sustainable agriculture development. The pixel-based image classification technique is often used to assess LULC changes (e.g., Maximum Likelihood Classification (MLC) and Support Vector Machine (SVM)). Object-based image analysis (OBIA) has also been widely used for LULC change detection. Therefore, the paper aims to compare the most common image classification methods for assessing LULC change in the agricultural arid of Tabuk using medium-resolution data from Landsat (1985-2015). Based exclusively on overall accuracy assessments, there was no advantage to preferring one image analysis method over another for the purposes of assessing LULC changes in arid environments using medium spatial resolution imagery. Visual interpretation, however, shows that the OBIA provided more accurate and satisfying results. Both pixels-based and object-based results indicate their possible further applicability for assessing LULC changes and corresponding studies. The result suggests that OBIA has the potential to be an alternative method over pixel-based methods for assessing LULC information taken over the spatially mixed land cover of arid Saudi Arabia.