Image fusion as an important branch of image processing field, has become one of the hot research issue. For pixel level image fusion, multi-scale and multi-resolution decomposition has been widely used. The low-frequency sub-band fusion method based on regional energy is proposed for the problems of fusion image fuzzy and contrast reduction from average method. For high-frequency sub-band, for considering the coefficient itself, the impact of regional coefficient is also considered. The registered visible and infrared images are fused and the fused images are evaluated. Experimental results show that while retaining the image outline, the high-frequency edge and details of the images are also fully retained.