Analyzed the limitations of existing down-top attention model in the application of coal mine surveillance video,then proposed a new computation model of visual attention aiming at coal-mine surveillance video based on sequential scale space and multi-features.Different from the existing model,the expression of non-uniform sampling was based on the discrete structure of sequential scale space,and chose the modified Bessel function as the smooth kernel.About feature extraction,chose the motion conspicuity,wavelet package decomposition and gray intensity as measures of saliency,and adopted DOG(Difference of Gaussian) operator as the generalized method.Finally,a global saliency map for the interesting objects was formed.The experiment results show the flexibility and effectiveness of this model.