Abstract: In this paper,we propose a novel hierarchical classification algorithm based on feature selection and adaptive decision tree in SAR image classification.Firstly,Joint Boosting selects feature combination most suitable for each class;Secondly,a hierarchical classifier is searched adaptively using binary classifiers based on feature combination;finally,we perform SAR image study and inference based on feature selection results and adaptive decision tree,leading to automatic classification.Experimental results on the first batch of PolInSAR data prove the proposed approach's efficiency.