Abstract: The insulation of overhead line in 10 kV distribution network is unsatisfied and lightning tripping accidents occur frequently in thunderstorm weather. Due to the complexity of large-scale line model construction, it is difficult to realize the risk assessment at the grid level based on physical model, such as lightning tripping rate. Besides, due to the lack of monitoring devices in the distribution network, the data quality is undesirable, which seriously affects the calculation accuracy of the model. In this paper, a lightning risk assessment method for 10 kV distribution network was proposed. Based on multi-dimensional data such as line operation, and meteorological environment information, the lightning assessment for large-scale distribution network lines was carried out under the condition of low quality data. Setting the geographical grid as evaluation unit, this paper extracted ten lightning-strike risk indexes from three dimensions, including meteorological, corridor environment, and characteristic parameters of line. Then TOPSIS algorithm improved by weighted Mahalanobis distance was used to build the lightning risk assessment system of 10 kV distribution network, and the risk was scaled. Finally, risk assessment results were verified by historical trip statistics in Guangzhou 10 kV distribution network, and the effectiveness of the assessment method was proved.