Abstract: In response to the problem that in existing base station indoor localization algorithm the selection of localization base stations and the unreasonable setting of weights may lead to low localization accuracy, this paper proposes an improved weighted centroid indoor localization algorithm based on received signal strength indicator (RSSI) in the 5G environment. The algorithm obtains the distance from five known base stations to the point to be located by RSSI ranging, and then makes a circle with the known base station location as the centre. For the intersecting pentagonal area, any three vertices are taken to form a triangle. The triangle's centroid coordinates are calculated by setting the appropriate weights according to the type of base station and the distance to the point to be located, and then a maximum likelihood estimation is made based on the ten triangle's centroid coordinates to obtain the final location point. The simulation results show that the indoor localization accuracy of this algorithm is significantly improved when compared with the normal and weighted centroid algorithms in both sparse and dense base station environments.