This paper introduces the Scale-recurrent Network (SRN-DeblurNet) for deep image deblurring, a technique that addresses the "coarse-to-fine" scheme by progressively restoring sharp images at different resolutions. The proposed method is designed to be simpler, more parameter-efficient, and easier to train compared to existing learning-based approaches. The SRN-DeblurNet incorporates a scale-recurrent structure, where network weights are shared across scales, reducing the number of parameters and improving stability. Additionally, it uses an Encoder-decoder ResBlock network, which enhances receptive field size and training efficiency. The method is evaluated on large-scale datasets with complex motion, demonstrating superior performance in both quantitative and qualitative metrics. The results show that SRN-DeblurNet outperforms state-of-the-art methods in terms of image quality and restoration accuracy.This paper introduces the Scale-recurrent Network (SRN-DeblurNet) for deep image deblurring, a technique that addresses the "coarse-to-fine" scheme by progressively restoring sharp images at different resolutions. The proposed method is designed to be simpler, more parameter-efficient, and easier to train compared to existing learning-based approaches. The SRN-DeblurNet incorporates a scale-recurrent structure, where network weights are shared across scales, reducing the number of parameters and improving stability. Additionally, it uses an Encoder-decoder ResBlock network, which enhances receptive field size and training efficiency. The method is evaluated on large-scale datasets with complex motion, demonstrating superior performance in both quantitative and qualitative metrics. The results show that SRN-DeblurNet outperforms state-of-the-art methods in terms of image quality and restoration accuracy.