In order to improve the performance of high resolution range profile target recognition in low SNR, a robust target recognition method based on extended sparse representation is proposed. In this method, the local and global features of the target high-resolution range profile are extracted by the extended sparse representation. In the training phase, support vector theory and dictionary learning principle are used to optimize the feature extraction dictionary to improve the separability of feature vector. In the test stage, the model matching method of factor analysis is used to optimize the de noise dictionary, so as to effectively suppress the noise and ensure the noise robustness of the target recognition system. The performance of the method is tested with the measured data. The results show that the method can effectively recover the high-resolution range profile of the target under the condition of low SNR, and achieve high recognition accuracy.