The invention provides a nonnegative sparse matrix and hyperspherical color transformation-based image fusion method, and relates to the field of image fusion. The method comprises the following steps of: jointly extracting a brightness component of a full-color image and a multi-spectrum image by using a nonnegative matrix decomposition algorithm; carrying out modeling calculation and adjustment on the bright components; and fusing the images by using hyperspherical color transformation so as to obtain a fusion result image. According to the method, the full-color image and the multi-spectrum image are combined to extract an I component by utilizing an NMF algorithm, so that the brightness component extraction precision is improved while the problem of bad spectrum matching attribute is solved; and through adjusting the I component by utilizing the full-color image and a filtered image thereof, space detail information is better blended, so that spectrum distortion is prevented from the greatest extent, the fusion result images of novel satellites are greatly improved in the aspects of space detail information blending and spectrum characteristic keeping when being compared with the existing algorithm, and the subjective evaluations and the objective analysis results can be consistent.