Abstract: Space-based infrared small target detection is a key research direction in computer vision and remote sensing, with its core task being the accurate identification of dim, small, and textureless targets within complex infrared images captured under space-based conditions. In such scenarios, the technology must address challenges including space background clutter, atmospheric radiation interference, uncertain target motion trajectories, and limited onboard computational resources—placing high demands on the real-time performance, sensitivity, and robustness of detection algorithms. This paper systematically reviews the unique difficulties faced by space-based infrared small target detection, examines the applicability and limitations of classical methods in spaceborne environments, and focuses on recent advances in deep learning-based approaches. Key areas covered include single-frame and multi-frame detection frameworks, the construction of annotated datasets for space targets, lightweight network architecture design for satellite platforms, and issues related to onboard deployment and hardware-software co-optimization. Furthermore, the article discusses emerging trends in large-scale infrared detection models tailored for space applications, such as cross-modal fusion, onboard incremental learning, and collaborative star-ground inference systems. It aims to offer a comprehensive and forward-looking reference for researchers working in the field of space-based infrared surveillance and tracking.