Land cover observation based on remote sensing data demands robust classification techniques which give the precise complex land cover mapping. Scientists and researchers made great efforts in improving classification accuracy considerably. The aim of this paper is to show outcomes gained from the RF classifier and decision tree and to compare their effectiveness with the SVMs technique. The mentioned techniques are applied over the imagery we have captured with six different classes of ROI (Region Of Interest) images including unknown range. Results indicated that the performance of the random forest classifier outperforms the decision tree and SVMs techniques performance in terms of the number of mis-classifications instances and the classification accuracy with an overall accuracy of 86%, while the decision tree accuracy is 67%, and the SVMs accuracy is 56%, respectively.