Abstract: Aiming at the problem of high cost and low precision of three-dimensional localization of power inspection robots in indoor environment, this paper proposes a 5G indoor three-dimensional fingerprint localization algorithm for power inspection with 5G signal characteristics and multi-signal classification algorithm. Firstly, a three-dimensional channel model is constructed to reduce the deviation of positioning signal parameters caused by multi-path interference. Secondly, the multi-path positioning signal optimization algorithm is proposed to overcome the phase deviation caused by random noise, so as to reduce the hardware and software cost of positioning decision. Simulation results show that the proposed algorithm achieves good performance and balance in two aspects of three-dimensional indoor positioning accuracy and positioning cost.