Low-altitude target detection has wide applicable and significantly contributes to the economic growth of areas at low-altitude. Although significant advancements have been achieved in detecting targets, there is still a need to improve the real-time and accuracy required to detect small targets in some specific scenarios. In this paper puts forward Global Feature Enhancement Detection (GFE-Det) algorithm, which is an optimised single-stage target detection algorithm called YOLOX. It is designed for small target detection in low-altitude. Firstly, GFE-Det enhances small target features globally by using the YOLOX-s model and improving the CSPLayer layer in Neck. The Global Context(GC) module added to the CSPLayer is optimized for small target information extraction. Secondly, the GC is added before the feature fusion. This enhances model’s capacity to detect recognize and position the target of interest, thus hastening the detection process. Finally, the APD (Airplane and Parachute Datasets) dataset is constructed, consisting of 2410 images of UAVs and parachute targets at low-altitude. The GFE-Det approach presented in this paper achieves accuracy of 91.53% and 11.5% improvement in detection speed versus YOLOX-s on the APD dataset.