Imaging degradation widely exists in remote sensing images at different levels. Instances are studied to show that the influence of imaging degradation varies with quality of imaging system by simulating remote sensing process on the basis of imaging model. It appears that it is important to extract information from remote sensing imagery by a method under the consideration of image degradation. The experimental results show that the support vector machine algorithm can reduce the effect of nonlinear mixture in pixel or among pixels and improve classification accuracy, especially by combing support vector machine algorithm and image restoration method.