Abstract: The sparse K-SVD algorithm is a method that can produce a dictionary combined the structure of the analytical dictionaries and the adaption of the trained dictionaries effectively, resulting a dictionary that has a good sparse representation capability.Theoretically,this method can achieve better result than other methods.In order to further improve the efficiency of the image fusion method,we propose a novel image fusion method based on the sparse K-SVD dictionary.Finally,the experimental results prove the effectiveness of the algorithm.