This paper introduces a novel method for generating artistic portraits from face photos using feature disentanglement and reconstruction. The approach combines six key components: a U-Net-based image generator, an image discriminator, a feature-disentanglement module, a feature-reconstruction module, a U-Net-based information generator, and a cross-modal fusion module. The feature-disentanglement module uses wavelet transforms to separate content and style features, while the feature-reconstruction module merges these features with contextual information to generate realistic and diverse artistic portraits. The cross-modal fusion module integrates the reconstructed feature map with the image generator's output to produce the final artistic portrait. The method is evaluated on the APDrawing dataset, achieving a significantly lower Fréchet Inception Distance (FID) score compared to existing methods, indicating superior visual quality. Ablation studies confirm the effectiveness of each component, particularly the feature-disentanglement and reconstruction modules, in enhancing artistic quality. The method demonstrates strong performance in generating realistic and diverse artistic portraits, outperforming state-of-the-art techniques in terms of visual quality and diversity. The results highlight the effectiveness of the proposed method in balancing realism, structural similarity, and diversity in artistic portraits.This paper introduces a novel method for generating artistic portraits from face photos using feature disentanglement and reconstruction. The approach combines six key components: a U-Net-based image generator, an image discriminator, a feature-disentanglement module, a feature-reconstruction module, a U-Net-based information generator, and a cross-modal fusion module. The feature-disentanglement module uses wavelet transforms to separate content and style features, while the feature-reconstruction module merges these features with contextual information to generate realistic and diverse artistic portraits. The cross-modal fusion module integrates the reconstructed feature map with the image generator's output to produce the final artistic portrait. The method is evaluated on the APDrawing dataset, achieving a significantly lower Fréchet Inception Distance (FID) score compared to existing methods, indicating superior visual quality. Ablation studies confirm the effectiveness of each component, particularly the feature-disentanglement and reconstruction modules, in enhancing artistic quality. The method demonstrates strong performance in generating realistic and diverse artistic portraits, outperforming state-of-the-art techniques in terms of visual quality and diversity. The results highlight the effectiveness of the proposed method in balancing realism, structural similarity, and diversity in artistic portraits.