Under poor lighting conditions, by using the properties of infrared and visual image, based on wavelet transform a fusion algorithm for the infrared and visual image in the same scene is proposed. This fusion algorithm makes image edge in highlight and remains the spatial details. Firstly, wavelet is decomposed. Secondly, a fusion rule of correlation coefficient combining with weighting and selection is used for the low-frequency component which reflects approximate contents. The fusion method for regional average gradient combining with weighting and selection is used for the high-frequency component which reflects detail features of image. Finally, through an inverse transform of wavelet, the fusion image is reconstructed. Experimental results prove that by using this algorithm, the mutual information between the images can be fused organically, the image resolution is raised, the details of fusion image are enhanced, the edge information is reappeared better, and object from infrared image is showed. Strong support for the follow-up target identification, follow-up information analysis and extractive ability of the images is provided.