Building suitable image fusion techniques for remote sensing application is an emerging field of research. Though there exist quite a few algorithms in this domain, but still there is a scope of improvement in terms of quality of the fused image and reduction in the complexity of the fusion algorithms. In this paper, we have proposed a new adaptive fusion methodology, which is a modified form of the principle component analysis (PCA) technique based on a window technique. Our proposed method gives higher fusion quality compared to some of the existing standard methods, in terms of image quality and promises to be less complex. For our experiment, we have used the high spatial resolution panchromatic (PAN) image and the multispectral (MS) image, as available from remote sensing satellites such as SPOT5. © Springer International Publishing Switzerland 2014.