CU-Mamba is a novel model for image restoration that integrates a dual State Space Model (SSM) framework into the U-Net architecture. It combines a Spatial SSM module for global context encoding and a Channel SSM component to preserve channel correlation features, both operating with linear computational complexity relative to the feature map size. The model effectively captures long-range dependencies and preserves channel-wise correlation, leading to improved performance in image restoration tasks. Experimental results show that CU-Mamba outperforms existing state-of-the-art methods in image denoising and deblurring tasks, achieving higher PSNR and SSIM scores while maintaining lower computational costs. The model's dual SSM blocks enable a balance between spatial and channel context understanding, enhancing the quality and accuracy of restored images. The work highlights the importance of integrating both spatial and channel contexts in image restoration and offers a new perspective on the U-Net architecture.CU-Mamba is a novel model for image restoration that integrates a dual State Space Model (SSM) framework into the U-Net architecture. It combines a Spatial SSM module for global context encoding and a Channel SSM component to preserve channel correlation features, both operating with linear computational complexity relative to the feature map size. The model effectively captures long-range dependencies and preserves channel-wise correlation, leading to improved performance in image restoration tasks. Experimental results show that CU-Mamba outperforms existing state-of-the-art methods in image denoising and deblurring tasks, achieving higher PSNR and SSIM scores while maintaining lower computational costs. The model's dual SSM blocks enable a balance between spatial and channel context understanding, enhancing the quality and accuracy of restored images. The work highlights the importance of integrating both spatial and channel contexts in image restoration and offers a new perspective on the U-Net architecture.