Numerous paintings in art galleries remain unattributed, with their creators yet to be identified. This challenge seeks to foster the development of a model capable of identifying the artist of a painting based on its digitized image. This paper introduces an efficient automatic artist identification system that utilizes texture features and color moments. The proposed system comprises several stages. Utilizing the existing Painting 91 dataset requires pre-processing. Next, we extract and normalize the local binary pattern (LBP), gray level co-occurrence matrix (GLCM), and color moments. Advanced feature selection techniques, including SHAP (SHapley Additive exPlanations), are compared with the Sequential Forward Selection method to evaluate their effectiveness in reducing the feature space and identifying significant features capable of distinguishing paintings by different artists. The classification accuracy for artists such as Jackson Pollock and Camille Pissarro exceeds 70%, performing competitively with state-of-the-art methods.