Steganalysis of spread spectrum image steganography is an important aspect in the domain of information hiding research.An image statistical distribution model based on high-order Markov chain was proposed.By a comparison on the effect of n-order Markov chain achieved by several general image scanning methods,Hilbert scanning method was used to construct the proposed model.After that,a n-order statistic security measure for digital image steganography based on the model was proposed,and that it is bounded was proved.Finally,after the effect of spread spectrum steganography on empirical matrix of the proposed model was studied,image statistical features which based on the proposed model were extracted,and several common spread spectrum steganography algorithms were analyzed using support vector machine.Experiment shows the active effects of steganalysis of spread spectrum image steganography based on the proposed model is obvious,furthermore,the steganalysis accuracy is improved as the model order increased.