摘要： In order to reduce the correlation between adjacent pixels in a plaintext image, and to solve the small key space problem of image encryption algorithms when a low-dimension chaotic map is used, a new encryption method for color images is proposed based on two-dimensional variational mode decomposition (2D-VMD) combined with eight-dimensional (8D) hyper-chaotic systems. 2D-VMD decomposes R, G and B components of a color image respectively. The 8D hyper-chaotic system is constructed by means of variable coupling. After improving and combining the original sequences obtained by iterating the system, two groups of key sequences associated with the plain-image are obtained. One group is used to scramble each mode image obtained by 2D-VMD; the other is used to replace the pixel values of the scrambled mode images. For different mode images, different key sequences are adopted. The encrypted images whose number is equal to the number of the modes can then be obtained. The experimental results show that, compared with the existing methods, the correlation coefficients between the pixels in the spatial domain of the plain-image can be reduced by 2D-VMD, as it’s more difficult to crack. In addition, the encryption method has better statistical and differential characteristics, as well as large enough key space, and better plain-image sensitivity.