In order to make up for the outstanding problems such as insufficient data integration and single methods in the risk assessment of open pit mines,a dual-channel multi-index assessment model based on GIS and census platform was developed to deconstruct the risk dual structure framework of "probability of occurrence - severity of consequences". A four-dimension evaluation index system including historical accidents (Q1), disaster defense (Q3), casualties (M) and economic losses (E) should be formed, and a four-level risk quantification and classification standard should be established. Taking the open pit mine of Sangzhi County as the research object, we know that 93.3% of the risk (Ⅲ grade) is more likely to occur in the open pit mine, and the protection against natural disasters does not meet the requirements. The value of economic loss was only 4 (≥￥10000), and only one open-pit mine reached the level of Ⅱ. In the end, 90% of the risk levels of open pit mines failed to meet the requirements of no historical accidents, natural disaster prevention, low casualties and low economic losses at the same time. The spatial heterogeneity of risk was accurately revealed with the help of visualization maps, and the evaluation results were 99% consistent with the actual disaster records. The integration of multi-source heterogeneous data and spatial modeling technology can significantly improve the accuracy and reliability of risk identification, and provide a new technical paradigm for the construction of a new disaster prevention system of "data-driven, intelligent assessment and spatial decision-making" in mining areas, which has important theoretical significance and practical value.