Spec-Gaussian introduces an anisotropic spherical Gaussian (ASG) appearance field to enhance 3D Gaussian splatting (3D-GS) in modeling specular and anisotropic components. Traditional spherical harmonics (SH) struggle with high-frequency details, leading to poor rendering of scenes with reflections and anisotropy. Spec-Gaussian replaces SH with ASG, enabling more accurate representation of these features without increasing the number of 3D Gaussians. A coarse-to-fine training strategy is introduced to improve learning efficiency and reduce floaters in real-world scenes. Additionally, anchor-based 3D Gaussians are used to reduce storage overhead and accelerate rendering. The ASG appearance field allows 3D-GS to better capture high-frequency details in complex scenes, improving rendering quality. Experiments show that Spec-Gaussian outperforms existing methods in rendering quality and efficiency, particularly in handling specular and anisotropic surfaces. The method is evaluated on both synthetic and real-world datasets, demonstrating its effectiveness in various scenarios. The approach also includes a novel feature decoupling MLP to separate diffuse and specular components, enhancing the ability to model complex optical phenomena. Overall, Spec-Gaussian significantly improves the performance of 3D-GS in rendering scenes with high-frequency details and complex geometries.Spec-Gaussian introduces an anisotropic spherical Gaussian (ASG) appearance field to enhance 3D Gaussian splatting (3D-GS) in modeling specular and anisotropic components. Traditional spherical harmonics (SH) struggle with high-frequency details, leading to poor rendering of scenes with reflections and anisotropy. Spec-Gaussian replaces SH with ASG, enabling more accurate representation of these features without increasing the number of 3D Gaussians. A coarse-to-fine training strategy is introduced to improve learning efficiency and reduce floaters in real-world scenes. Additionally, anchor-based 3D Gaussians are used to reduce storage overhead and accelerate rendering. The ASG appearance field allows 3D-GS to better capture high-frequency details in complex scenes, improving rendering quality. Experiments show that Spec-Gaussian outperforms existing methods in rendering quality and efficiency, particularly in handling specular and anisotropic surfaces. The method is evaluated on both synthetic and real-world datasets, demonstrating its effectiveness in various scenarios. The approach also includes a novel feature decoupling MLP to separate diffuse and specular components, enhancing the ability to model complex optical phenomena. Overall, Spec-Gaussian significantly improves the performance of 3D-GS in rendering scenes with high-frequency details and complex geometries.