A kind of image fusion system which contained short wave infrared detection, medium wave infrared detection and laser ranging device had been developed. The system realized three functions which were remote scene monitoring, remote target ranging and detecting the laser beam at 1.06 and 1.57 μm bands during all day and through the fog. In order to ensure the fusion image quality of system and improve the registration accuracy, an optical axis parallelism correction method to realize the coarse registration between the short wave infrared image and medium wave infrared image has been proposed; a new iteration optimization method which found the optimal register parameters to get the best fusion results has been put forward. This method used definition index of the fusion image interesting region to improve the register process, and used particle swarm optimization algorithm to solve joint optimization problem and to realize the accurate registration. The experimental results demonstrated the potential of the system for applications like remote scene monitoring and battlefield situational awareness.