In this paper, we propose an effective feature extraction algorithm, called Multi-Subregion based Correlation Filter Bank (MS-CFB), for robust face recognition. MS-CFB combines the benefits of global-based and local-based feature extraction algorithms, where multiple correlation filters correspond- ing to different face subregions are jointly designed to optimize the overall correlation outputs. Furthermore, we reduce the computational complexi- ty of MS-CFB by designing the correlation filter bank in the spatial domain and improve its generalization capability by capitalizing on the unconstrained form during the filter bank design process. MS-CFB not only takes the d- ifferences among face subregions into account, but also effectively exploits the discriminative information in face subregions. Experimental results on various public face databases demonstrate that the proposed algorithm pro- vides a better feature representation for classification and achieves higher recognition rates compared with several state-of-the-art algorithms.