The PDE-based diffusion model is very effective in reducing noise and preserving edges, which are the major problems in image processing. This paper aims to propose a time-fractional diffusion model to remove additive noise from noisy images while keeping important edges clear. The proposed model uses both spatial derivatives and a time-fractional derivative. This fractional order helps control the diffusion process more effectively than the classical model. The model is discretized using a finite difference method for numerical implementation. The outcomes of the fractional model are evaluated using peak signal-to- noise ratio (PSNR), and the results show that the model improves noise removal while preserving image edges.