Abstract: In order to realize the spatiotemporal dynamic observation of smart grid and optimize the configuration of metering equipment, a spatiotemporal dynamic positioning model of smart grid metering equipment based on random geometry is proposed. Stochastic geometry modeling tool described the spatial properties of power grid through bus, and also captured the time expansion of power grid. The K-means clustering algorithm was used to divide the power grid into sub networks and cluster them. A constrained finite time domain Markov decision process algorithm was used to solve the equipment optimal configuration problem under budget constraints. Simulation results show the effectiveness of the proposed method.