In this study, we introduce a novel image fusion model that incorporates a fidelity term on a subregion alongside a nonlocal version of the Poisson equation. The model leverages nonlocal differential operators to exploit similarities between distant pixels and regions. To establish the existence and uniqueness of the solution to the evolution equation, we apply Banach’s fixed-point theorem. Experimental results demonstrate that our proposed model achieves state-of-the-art performance in terms of visual quality.