Image fusion can integrate redundant and complementary information from multiple image sources,and the comprehensiveness of image understanding can be increased,and a more accurate and reliable image can be obtained for the same scene. Fusion rules of traditional wavelet image fusion are to process low-frequency and high-frequency coefficients respectively,but it ignored the relation between the two. The contrast is used to connect low-frequency coefficients process with high-frequency coefficients process,and it is as the criterion to select the coefficient for image fusion. Experimental simulation results show that the algorithm can preserve the useful information from multiple source images,and it is an effective image fusion algorithm.