Abstract: Pipeline blockage is one of the main failure accidents of filling pipeline which seriously hinders the normal operation of filling. To accurately predict the risk of blockage,which based on the uncertainty and fuzziness of filling pipeline blockage,a multi-dimensional cloud model was established. Eleven elements,including the stowing gradient,were selected as the variables of the cloud model. The certain degree of each indicator of different risk levels was calculated according to the cloud theory. Taking the multi-index interaction of pipeline blockage into consideration,we introduce the improved CRITIC method to process the index data,and obtained the comprehensive determinacy by combining the multi-dimensional cloud model,so as to determine the risk level of filling pipeline blockage. The paper studied the case of four domestic mining enterprises,the results showed that the wear risk grades of Pipeline blockage in Longshou Mine,Hedong Gold Mine,Dahongshan Copper Mine and Xincheng Gold Mine were Ⅱ,Ⅲ,Ⅱ and Ⅲ respectively. Compared with fuzzy mathematics,unascertained measure theory,one-dimensional cloud model and SPA-AHP,the pipeline blockage risk grades of Dahongshan Copper Mine obtained by this method and other methods are Ⅱ,Ⅳ,Ⅲ,Ⅲ and Ⅱ respectively,the calculation model is more rigorous to the assessment results and can ensure safety in practice on the premise of satisfying the accuracy of assessment results. It also provides a new method for the risk prediction of filling pipeline blockage.