Fusion method of subdivision band images of mid-wave infrared is presented based on both morphology processing and support value transform (SVT). Local process of two subdivision band MWIR images is implemented with erosion, dilation and weighted stacking method, in order to eliminate saturating regions in the first subdivision band image that are formed by sun irradiation. Fusion between local processed image and the second subdivision band image is achieved by SVT, in order to combine detailed information of two images and to get legible results. The experimental results show that the saturating regions of image obtained by proposed method, compared with image fused by wavelet packet transform, are removed effectively. The sun effect parameter decreased by 60.05%, local standard deviation and local coarseness increased by 2.59% and 3.39% respectively, and elapsed time decreased by 66.77%. The validity of fusion method proposed is proved.