Abstract: In order to repair large numbers of Dunhuang Murals with intricate damaged area and unbelievable vague information,the Dempster-Shafer evidence theory and its data fusion,one algorithm for repairing images,is presented in this paper,which improves the priority function of repair area filling algorithm.It is proposed on the basis of being taking full consideration of information complexity of the Dunhuang Murals,visual effects of mural inpainting,the reasonableness of repair being easily influenced by pixel repair order,and the other factors.Firstly the Dempster-Shafer evidence theory and its data fusion are used to transfer clear factors and data factors in the edge of repaired area separately to their own basic probability value functions,then a new priority function is obtained through fusing data,and finally we will see the perfectly repaired Murals.From the point of view of the visual psychology,the experimental results show that the repairing effect is improved and the effectiveness of the proposed algorithm is much better than those of others.For example,it gets rid of the disadvantage of Criminisi algorithm and its improved algorithm.