This paper focuses on the fusion of infrared and visible images based on the theory of sparse representation,and a fusion rule of sparse coefficients is proposed.The approach can be divided into three parts:over-complete dictionary,the algorithm of sparse vector approximation and the fusion rule.Firstly,through dictionary learning to all the patches of the visible image and infrared image to be fused by use of K-means Singular Value Decomposition (K-SVD),the over-complete dictionary for sparse vector calculation is obtained.Secondly,the sparse vector is approximated by orthogonal matching pursuit.Thirdly,the fusion rule based on absolute value of the maximum element of sparse vector is proposed,which is used for the sparse vector fusion of visible image with infrared image,and the fusion image is obtained.Experimental results show that the fusion result is obviously better than that of the method based on Maximum-L1-Norm.