Rain can make image blurry or unclear in various critical uses cases, single-image deraining often finds, difficult to precisely remove rain streaks due to the diverse range of rain densities and raindrop sizes. In autonomous real-time uses, with existing networks like U-Net, future decoupling reorganization, de-graded attention transformer, progressive multi-scale deraining network and dual-domain strip attention module finds low PSNR and SSIM. We proposed a novel Integrated Progressive Multi-scale Deraining Network (IPMDNet) to address peak signal to noise ratio (PSNR) enhancement, structural similarity index matrix (SSIM) improvement, and low loss of original data. The network supports a U-Net architecture combined with additional convolution filters to effectively handle the non-uniformity, of rain densities and the range of raindrop scale. A multi-scale strategy is used to capture rain streaks across various scales. Channel attention blocks are integrated to enhance the precision of rain streak removal and refine feature extraction. Furthermore, to improve deraining performance, we present a “Multi-scale Feature Integration” (MSFI) that refines and aggregates the features across several network stages. IPMDNet exceeds existing state-of-the-art techniques in many kinds of synthetic and real-world dataset tests, achieving better results in both quantitative and qualitative evolutions.