Image fusion is a very useful method to combine the spectral information from multispectral images with the high spatial resolution panchromatic images to improve the spatial resolution of multispectral images. The fast Intensity-Hue-Saturation (IHS) fusion technique proposed by Tu is based on color systems transformation between RGB and IHS color spaces . In 2006, Choi improved this method by introducing a tradeoff parameter which can control spatial resolution of fusion result . But Choi's method only uses single parameter for all pixels. In this study, we use the correlation of each pixel in multispectral image to create a transform, and assign tradeoff parameter for each pixel by a linear transform to reach minimum mean squared error between the fused image and the origin one.