Currently available UHF partial discharge（PD） localization technology is mainly based on the time delay algorithm. To obtain an accurate time delay between each channel of PD pulse requires a very high sampling rate

thus expensive hardware costs are needed and it is hard to achieve wide application. We proposed an RSSI-fingerprinting-based PD localization methodology

which possesses the advantages of low hardware cost and easy fulfillment in RSSI and features of being well adaptive to complex space environment of fingerprinting. The proposed method includes two stages: the RSSI fingerprinting setup during offline stage and the online PD localization by particle swarm optimization（PSO） based BP neural network（PSO-BP）. The field tests show that the mean error of PD localization by the proposed method is 0.484 m

and 87% of errors are less than 1 m while the corresponding variance is small

verifying the accuracy and stability of proposed novel PD localization method.