Addressing the challenging problem of detecting small targets in space, an infrared small target detection algorithm that incorporates foreground moving targets and deep learning is introduced. By running the foreground motion target detection algorithm and the deep learning YOLOv5 target detection algorithm, the regions where the motion target is located are obtained respectively. Then, the suspicious region of motion small targets is focused on and framed through region comparison. To verify the effectiveness of the algorithm, an infrared image dataset was constructed, and the experimental results demonstrate that the designed algorithm can effectively achieve intelligent detection of infrared small targets from kilometers away. This offers substantial support for practical applications in the fields of space exploration, deep space exploration, and security monitoring. It holds substantial application value for the integrated air and space monitoring and detection of small targets in space.