This paper proposes a multi-stage, feature-aware enhancement and restoration framework tailored for AI-generated traditional Chinese painting images. Addressing prevalent issues in current artificial intelligence generated content (AIGC) systems, including blurred brushwork, missing texture details, color distortion, and noise interference, the method integrates adaptive luminance correction, structure-preserving denoising, and multi-scale detail reconstruction. The framework operates in three sequential stages: (1) an adaptive gamma correction module that normalizes luminance distribution to enhance ink-wash clarity; (2) a novel texture-feature weighted bilateral filtering strategy that dynamically suppresses noise while preserving and enhancing stroke integrity and artistic style; and (3) a multi-scale wavelet fusion module that reconstructs fine textures and hierarchical details. Comprehensive experiments demonstrate that the proposed method significantly improves the clarity, structural integrity, and artistic expressiveness of AIGC-generated Chinese paintings. It outperforms both conventional image processing methods and state-of-the-art learning-based models in objective metrics (PSNR, SSIM, LPIPS) and subjective visual fidelity assessments. Our framework is training-free with zero learnable parameters and real-time inference (0.85 s per 512 × 512 image), making it lightweight and practical for deployment. This work provides a practical and effective solution for enhancing the quality of AI-generated art, facilitating its application in digital cultural heritage.