Abstract: In order to meet the requirement of power grid company for position information of metering terminal, this paper adopts the method of positioning metering terminal by using wireless transmission network, and then studies the positioning algorithm. The basic positioning algorithm uses the typical three-point positioning algorithm, and uses the signal intensity and position information of the base station to locate the terminal. At the same time, aiming at the non-line-of-sight error in the process of signal transmission, BP artificial neural network algorithm is used to adjust the weight value according to the environment to reduce the error. The Matlab simulation results show that the distance trained by BP neural network is closer to the expected value. At the same time, in the process of applying the training model, the humidity information of the application environment is collected and the parameters are adjusted to make the results more real-time.