Traditional methods for camouflage effectiveness evaluation often rely on manually extracting multiple features and calculating the overall camouflage performance, which cannot provide a quick and accurate assessment of the degree of camouflage in an image. Furthermore, in the field of deep learning, the ranking of camouflage objects involves measuring the degree of camouflage, but the effectiveness of image segmentation in this task can affect the classification results. To address these issues, this paper puts forward a kind of camouflage effect evaluation model based on attention mechanism, the camouflage ranking task by subtracting the segmentation steps to optimize the task mode, and carefully design features similarity calculation module, divided the corresponding camouflage data set, using local modeling advantage of attention mechanism, we measure the degree of camouflage more effectively and directly. Experimental results demonstrate that the Spearman rank order correlation coefficient (SROCC) and Pearson linear correlation coefficient (PLCC) of the model on the redivided CAM-LDR dataset are improved by 70.9% and 57.9%, respectively, compared with the current best model. In addition, the root mean square error (RMSE) was reduced by 9.25%. Our work is a new and effective attempt to extend disguised assessment in the field of deep learning.