Image fusion is a technique of combining complementary information from multiple images originated from single or different sources to form single fused image. This fused image has better quality compared to the original input images. In this paper a new fusion technique based on independent component analysis (ICA) and IHS transformation is proposed. A comparison of this new technique with PCA, IHS, and ICA-based fusion techniques is given. Ikonos-2 data are used to test the new and other three techniques. Four evaluation criteria are used for quantitative assessment the quality of the obtained fused images. To assess the spectral quality the spectral discrepancy ( Dk) and correlation coefficients (cc ) are used. For spatial quality assessment high pass correlation coefficients (HPCC ) and information entropy ( EN) are used. The results of the new technique show higher performance compared to the other techniques.