In order to improve the fusion image quality and target detection performance of infrared and visible image fusion system, an image fusion reconnaissance system was designed which used moving target of infrared video and visual to moniter the scene. This system realized the infrared image segmentation algorithm based on human visual characteristics and weighted fusion algorithm based on target′s characteristics in the hardware platform. Also, the system contained the circuit of infrared moving object segmentation and fusion processing circuit. Firstly, image segmentation circuit to extract the moving target of infrared video was designed by FPGA. Secondly, the image which only contained moving target was input to fusion processing circuit consisting of DSP. Last, a series of fusion images which had the characteristic of target form prominent and background details clear were furnished. Through the experiment, this system can both improve the target detection performance and processing speed. The image fusion evaluation increased more than 90%, and some indexes increased more than 7 times.