In this paper, we propose a pattern classification method for remote sensing data using both a neural network and knowledge-based processing. A neural network has the ability to recognize complex patterns, and classifies them to one of the classes. However,the neural network might produce misclassification. A knowledge-based system which uses human geographical knowledge improves the classification results, compared with a conventional statistical method. The disadvantage of using a knowledge-based system is that it needs a large amount of knowledge to classify the data correctly. We propose a pattern classification method that integrates the advantages of both the neural network and knowledge-based system. The proposed system is divided into two subsystems which consist of recognition and error correction. We use the neural network for classification and the knowledge-based system for correcting misclassification created by the neural network. Experimental results are shown to illustrate the performance of the proposed system.