This paper introduces a novel dictionary learning (DL) method called Fisher Discrimination Dictionary Learning (FDDL) to enhance pattern classification performance in sparse representation-based classification. FDDL learns a structured dictionary where each dictionary atom corresponds to a class label, and the reconstruction error after sparse coding is used for classification. The Fisher discrimination criterion is imposed on the coding coefficients to ensure small within-class scatter and large between-class scatter. The proposed method is evaluated on benchmark image databases, including face recognition, digit recognition, and gender classification, showing superior performance compared to existing methods. FDDL achieves higher classification accuracy with a smaller dictionary size, demonstrating its effectiveness and efficiency in various pattern recognition tasks.This paper introduces a novel dictionary learning (DL) method called Fisher Discrimination Dictionary Learning (FDDL) to enhance pattern classification performance in sparse representation-based classification. FDDL learns a structured dictionary where each dictionary atom corresponds to a class label, and the reconstruction error after sparse coding is used for classification. The Fisher discrimination criterion is imposed on the coding coefficients to ensure small within-class scatter and large between-class scatter. The proposed method is evaluated on benchmark image databases, including face recognition, digit recognition, and gender classification, showing superior performance compared to existing methods. FDDL achieves higher classification accuracy with a smaller dictionary size, demonstrating its effectiveness and efficiency in various pattern recognition tasks.