In order to realize pixel-level image fusion,and integrate source im ages′ edge information into the final fusion image,we propose a new image fusion method based on a multiscale guided filter decomposition and a new fusion strategy based on guided filter.Derived from a local linear model,the guided filter generates the filtering output by considering the content of a guidance image.For the proposed image fu sion algorithm,we first decompose the source image by a multiscale edge-preserving filter,and a basic image and a series of detail images are obtained.Then,we adopt different fusion strategies for the basic image an d the detail images,respectively, and the proposed new fusion strategy is based on the guided filter.Finally,the final fusion image can be computed by adding the fused basic image and the fused detail images.Experiments show th at the presented image fusion method can retain the spectral information and detail information in the final f used image.We compare our fusion algorithm with many competitive image fusion techniques in terms of mutual infor mation (MI) and visual quality,and our method can achieve a higher normalized MI valu e,outperforms other competing methods by more than 0.5in MI value.In this work,the multiscale guided filter de composition is first introduced into image fusion problem,it can better reflect the image details,and provide more abundant info rmation for image fusion.