Aiming at low positioning accuracy and high computational complexity of the traditional indoor positioning method under large area and low fingerprint density, a new segmented high accuracy positioning algorithm based on linear interpolation and distribution overlapping is proposed. The algorithm uses the traditional nearest neighbor method for coarse positioning, and then updates the fingerprint database in the trusted area with linear interpolation method to increase the fingerprint density. Finally, fine positioning is realized by fingerprint matching method that based on distribution overlapping. Experimental results show that, even under low fingerprint density, the proposed method can get good positioning performance with high accuracy and suitable complexity.