Super-resolution reconstruction stands as a critical task within the domain of computer vision. To enhance texture information extraction and improve visual perception, we introduce the Multi-Scale Hybrid Attention Network (MSHA), a novel single-image super-resolution model engineered to augment image global processing capabilities while accentuating detail reconstruction. The MSHA architecture seamlessly integrates a Multi-Scale Feature Block (MFB) for comprehensive feature extraction and a Parallel Hybrid Attention Module (PHA) for refining detail reconstruction capabilities. Through extensive experimentation and comparative analyses against state-of-the-art models, we substantiate the superior performance of the MSHA network in producing high-quality super-resolution images. Our methodology effectively addresses the shortcomings of existing approaches by emphasizing multi-scale feature extraction and detail reconstruction, thereby significantly advancing the field of image super-resolution.