Abstract: Infrared images are widely used in fields such as military surveillance, medical imaging, and remote sensing. However, due to limitations in imaging devices and environmental factors, their resolution is typically low. Infrared image super-resolution reconstruction refers to the process of reconstructing high-resolution infrared images from low-resolution inputs using specific algorithms or models. This technique enhances image clarity and recognizability, thereby meeting the stringent requirements for infrared image quality across various application scenarios. This study systematically reviews infrared image super-resolution reconstruction techniques developed in recent years, encompassing both traditional and deep learning methods, with particular emphasis on deep learning strategies tailored to the characteristics of infrared images and their associated degradation models. The review covers three mainstream network architectures: Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), and Transformers. Through a comparative analysis of experimental results from infrared image super-resolution methods proposed in the Thermal Infrared Super-Resolution (TISR) competitions from 2020 to 2024, this study identifies the current limitations of existing research and outlines potential future directions. The study aims to provide a valuable reference for subsequent research in infrared image super-resolution reconstruction.