Most deep learning-based restoration methods tend to concentrate solely on a single type of degradation, aiming either at enhancing low-light images or performing deblurring. Although the performances of these restoration models are impressive, they cannot simultaneously correct these degradations. Therefore, this article introduces a U-shaped network architecture for simultaneous low-light image enhancement and deblurring. The proposed architecture includes three main components: an encoder block that progressively extracts multi-kernel features while reducing spatial resolution, a middle block that processes the most abstract feature representations, and a decoder block that reconstructs the enhanced and deblurred output through upsampling and feature fusion. The experimental results demonstrate the capability of the proposed network can obtain promising results in terms of objective and subjective performances.