The invention relates to a power grid disaster risk assessment method and device based on strong precipitation space-time distribution characteristics. Rainfall data of a specific area is adopted. Data such as power supply quantity, whole society power consumption, per capita domestic production total value (GDP), high-resolution terrains and water systems and the like are used. On the basis of comprehensively considering the risk of disaster-causing factors (rainstorm and short-time heavy rainfall), sensibility of hazard inducing environment (terrains and river systems), vulnerability of disaster-bearing bodies (power supply quantity and power consumption quantity) and disaster prevention and resistance (per capita GDP). By utilizing the zoning method provided by the invention, a plurality of high-risk areas of the power grid heavy rainfall disasters in a specific area can be identified, and scientific meteorological information can be provided for power grid design, maintenance and decision service.