GSEdit is a text-guided 3D object editing method based on Gaussian Splatting. It enables efficient editing of 3D objects' style and appearance without altering their main details, using a pretrained diffusion model and Gaussian Splatting to represent 3D scenes. The method optimizes the model while varying image supervision, ensuring consistency across different viewpoints. Compared to previous methods, GSEdit is more efficient, allowing 3D editing tasks to be completed in minutes on consumer hardware. The editing process is refined using the SDS loss, ensuring precision and accuracy. GSEdit can handle various input formats, including 3D meshes and Gaussians from generative models. The method is evaluated on multiple datasets, showing effective alterations of object shape and appearance while preserving coherence and detail. GSEdit's efficiency and coherence make it a significant contribution to 3D editing, with potential applications in 3D vision and graphics. Limitations include perspective bias and restricted spatial transformations due to reliance on the Instruct-Pix2Pix framework.GSEdit is a text-guided 3D object editing method based on Gaussian Splatting. It enables efficient editing of 3D objects' style and appearance without altering their main details, using a pretrained diffusion model and Gaussian Splatting to represent 3D scenes. The method optimizes the model while varying image supervision, ensuring consistency across different viewpoints. Compared to previous methods, GSEdit is more efficient, allowing 3D editing tasks to be completed in minutes on consumer hardware. The editing process is refined using the SDS loss, ensuring precision and accuracy. GSEdit can handle various input formats, including 3D meshes and Gaussians from generative models. The method is evaluated on multiple datasets, showing effective alterations of object shape and appearance while preserving coherence and detail. GSEdit's efficiency and coherence make it a significant contribution to 3D editing, with potential applications in 3D vision and graphics. Limitations include perspective bias and restricted spatial transformations due to reliance on the Instruct-Pix2Pix framework.