Aiming at the low accuracy of traditional indoor positioning algorithm, an indoor positioning algorithm based on Elman neural network was proposed. The Elman neural network is used to expand the fingerprint database interpolation, complete the fingerprint database, and reduce the positioning error, and the algorithm model is verified by experiments. The experimental results show that the average positioning error of the proposed algorithm is 4.6 cm in 0.8 m × 0.8 m×0.8 m environment, which meets the accuracy requirements of indoor positioning.