In order to improve the accuracy of indoor visible light positioning, a single light-emitting diode (LED) lamp indoor po-sitioning algorithm based on double back propagation(BP) neural network is proposed. The algorithm first uses the BP neural net-work to determine the rough position range of the target in the positioning area, and then uses the position range as a limiting condition to use the BP neural network again to achieve more accurate positioning. The indoor positioning system uses a single LED lamp as the transmitter and three horizontal photodetectors as the receiver to receive optical power, which avoids the inter-symbol interference caused by using multiple LED lamps during positioning. The simulation results show that in the positioning area of 3 m×3 m×3.5 m, the average positioning accuracy of the algorithm in the paper can reach 0.0042 m, which is higher than the traditional indoor visible light positioning algorithm.