The method of classifying and sampling weighted calculation of the distribution network voltage qualification rate has the problems of using dots and planes, which cannot reflect the local range and time period characteristics. Multi-source data such as DSCADA system, marketing management system and GIS are integrated in this paper. A hierarchical aggregation of household-transformer-line-station is proposed to characterize the spatial characteristics of the distribution network voltage. Based on the high-frequency sampling information, it is proposed to calculate the qualified rate of the supply voltage of feeders and transformers in different periods, identify the warning period, and analyze the time characteristics of the local voltage, so that the time and space characteristics of the distribution network voltage can be accurately characterized. Calculations based on data from a distribution network in Guangxi show that the proposed method of voltage spatiotemporal characteristics can accurately reflect the voltage details of the distribution network, which realizes the sorting of the failure rate of the supply voltage of feeders and transformers, and identifies the period of low voltage.