The invention discloses a guiding significance image fusion method based on wavelet transform. The method comprises the steps that calculation of a significant atlas of a source image: according to the space frequency information of the image, the significant atlas of the source image is calculated; construction of a weight graph based on the significant atlas: firstly, single layer wavelet decomposition is carried out on the significant atlas of the source image and the corresponding image, then combined with a guidance wave filtering with a guidance gradient, the weight graph of a wavelet coefficient and a scale coefficient of each image is constructed; generation of a fusion image: according to the source image decomposed by wavelet and the corresponding weight atlas, the fusion image is formed by linear fusion, and then wavelet inverse transformation is carried out. The method can obtain a better effect on a multi-focus and multi-exposure image, has simple and convenient calculation, high efficiency and good realization, and can be applicable to a personal computer and transplanted to an embedded system conveniently. The method is applicable to computer vision and pattern identification, such as remote sensing image analysis, target tracking and target recognition.