Space target monitoring and detection is an important guarantee for maintaining the safety of the space environment, and space-based observation is one of the important observation methods. In space-based observation images, the shape and size of stars and small spatial targets are similar, which affects the detection of small spatial targets. Moreover, the spatial environment is complex, and the noise in the image can also interfere with the detection of small spatial targets. A method for detecting small spatial targets in orbit based on trajectory prediction is proposed to address the above issues. Firstly, the image is preprocessed, including image filtering, which is used to remove bad pixels and reduce background influence, and threshold segmentation. Distortion error correction on the camera is performed, then coordinate system conversion is performed as well, and the star catalog is matched to remove most stars. Then the method collects multiple frames of images, performs trajectory correlation and prediction on candidate targets, and determines the search range. Finally, the detection of small spatial targets is achieved. This detection method has the following advantages: Achieving high detection accuracy and low false alarm rate under low signal-to-noise ratio; Reducing search scope and improving real-time detection capability in orbit; Having the ability to detect discontinuous occurrences of targets on continuous frame images, adapting to the needs of target detection in more complex spatial environments. The detection ability of this detection method has been verified on a hardware system. The experimental results show that the algorithm has a high detection rate and a low false alarm rate when detecting low signal-to-noise ratio targets, which provides a reference for small target detection in orbit space.