The relatively recent national interest in the quality of the environment and in earth resources has created numerous areas of applications for pattern recognition technology. Among these areas, the automatic classification of remotely sensed data has received the greatest attention. Because of the large volumes of data generated by multispectrai scanners located abroad aircraft, satellites, and space stations, it has becomes necessary to search for automated means to process and analyzed this information. In our present work, computer programs are written and conducted to enhanced the appearance of the satellite multiband images and to classify their regions, automatically. The study has concerned with a type of satellite images, known as Thematic Mapper (TM) images, which usually provided in different "seven " bands. Our investigation have been devoted to analyze and classify available image regions, provided over the west of IRAQ country; namely RM4ADI region, An adaptive image enhancement technique, based on the modification of image histograms, is introduced. The adaptation in our presented method is that; unlike the histogram equalization enhancement method, number of quantized level can be decided in advanced. Two methods for classifying image regions are adopted and used for segmenting processed images. The first, named supervised classifier, based on minimum distance cretrion utilizes some statistical features (i.e. mean, standard deviation, minimum and maximum grey values), to classify different regions in an image. While the second method, named texture classifier, utilizes the features extracted from the co-occurrence matrices to discriminate between image regions (i.e. Contrast. Inverse moment. Entropy, and Correlation).