On the basis of analyzing immune learning mechanism, by modeling for image classification, we can solve the problem of remote sensing image classification by using the basic principles of the use of immune learning. We have realized a classification algorithm with a function of the immune learning. Classification algorithm divides each major category into a number of small categories and the antigen population evolutionary process of each category is considered separately, therefore the convergence time is greatly decreased. When classifying, we use a variety of different ways to discriminate and introduce artificial priori knowledge to improve the classification accuracy. The results show that the algorithm can be well applied in remote sensing image classification.