<p>In this paper, we elaborate a novel variational model for image denoising, particularly focusing on scenarios with high levels of speckle noise. Our approach integrates fractional-order regularizers to better preserve image edges, overcoming the limitations of traditional methods in handling small gradient variations. The main contributions of our work include the introduction of a non-convex attached term with a fractional-order gradient penalization to mitigate the staircasing effect. Additionally, we incorporate a controlled parameter to manage gradient magnitudes while preserving image contours and texture. Experimental results demonstrate the effectiveness and robustness of our proposed method, yielding competitive performance against state-of-the-art speckle noise removal techniques.</p>