Abstract: Remote sensing image fusion is an economical and effective approach for obtaining high-spatial-resolution hyperspectral images, capable of overcoming the limitations of single sensors. However, it involves an ill-posed inverse problem and is susceptible to noise contamination. To address these challenges, this paper proposes an image fusion model based on tensor ring decomposition, transforming the fusion process into the estimation of target image tensor ring factors. Low-dimensional subspace features are further utilized to achieve super-resolution reconstruction of high-dimensional data. First, the local similarity features of the tensor ring factors are exploited by constructing a multimode graph regularization term. Second, an approximation of global low-rank features in low-dimensional subspaces is obtained by introducing tensor nuclear norms for the truncated singular value decomposition of tensor ring factors. Finally, an efficient algorithm was designed to realize model optimization and solution. Experimental results on multiple datasets demonstrate that the proposed fusion model effectively enhances the spatial resolution of hyperspectral images while significantly suppressing noise.