This paper proposes a Scale-Recurrent Network (SRN-DeblurNet) for deep image deblurring. The method is designed to address the challenges of single-image deblurring, where the goal is to recover a sharp image from a motion- or focal-blurred input. The proposed SRN-DeblurNet follows a "coarse-to-fine" strategy, gradually restoring the image at different resolutions in a pyramid. It introduces a novel scale-recurrent network structure that shares weights across scales, reducing the number of trainable parameters and improving training efficiency. Additionally, it incorporates an encoder-decoder ResBlock network, which enhances the receptive field and improves performance on large-motion deblurring. The SRN-DeblurNet is evaluated on large-scale deblurring datasets with complex motion. Results show that the method outperforms state-of-the-art approaches in both quantitative and qualitative measures. The network structure is efficient, with a training time that is approximately one-quarter of that of existing methods. It also achieves higher quality results with significantly fewer parameters and faster testing times. The proposed SRN-DeblurNet uses a recurrent structure across multiple scales, where each scale's processing is based on the previous scale's output. This allows the network to capture useful information across different scales, improving the restoration of sharp images. The encoder-decoder ResBlock network further enhances the model's ability to handle complex motion by increasing the receptive field and incorporating residual learning blocks. The method is compared with several baseline models, including single-scale and multi-scale approaches, as well as other encoder-decoder structures. The results show that the SRN-DeblurNet outperforms these models in terms of PSNR and SSIM metrics. It also performs well on real-world blurred images, demonstrating its effectiveness in practical scenarios. The paper concludes that the proposed SRN-DeblurNet is a promising approach for image deblurring, with a simpler structure, better performance, and potential applications in other image processing tasks.This paper proposes a Scale-Recurrent Network (SRN-DeblurNet) for deep image deblurring. The method is designed to address the challenges of single-image deblurring, where the goal is to recover a sharp image from a motion- or focal-blurred input. The proposed SRN-DeblurNet follows a "coarse-to-fine" strategy, gradually restoring the image at different resolutions in a pyramid. It introduces a novel scale-recurrent network structure that shares weights across scales, reducing the number of trainable parameters and improving training efficiency. Additionally, it incorporates an encoder-decoder ResBlock network, which enhances the receptive field and improves performance on large-motion deblurring. The SRN-DeblurNet is evaluated on large-scale deblurring datasets with complex motion. Results show that the method outperforms state-of-the-art approaches in both quantitative and qualitative measures. The network structure is efficient, with a training time that is approximately one-quarter of that of existing methods. It also achieves higher quality results with significantly fewer parameters and faster testing times. The proposed SRN-DeblurNet uses a recurrent structure across multiple scales, where each scale's processing is based on the previous scale's output. This allows the network to capture useful information across different scales, improving the restoration of sharp images. The encoder-decoder ResBlock network further enhances the model's ability to handle complex motion by increasing the receptive field and incorporating residual learning blocks. The method is compared with several baseline models, including single-scale and multi-scale approaches, as well as other encoder-decoder structures. The results show that the SRN-DeblurNet outperforms these models in terms of PSNR and SSIM metrics. It also performs well on real-world blurred images, demonstrating its effectiveness in practical scenarios. The paper concludes that the proposed SRN-DeblurNet is a promising approach for image deblurring, with a simpler structure, better performance, and potential applications in other image processing tasks.