Aiming at the problem of low positioning accuracy to fault points in steel structures, an improved DV-Hop location algorithm is proposed, and a scheme of fault points detection based on the wireless sensor network technology is designed. This paper described the fault point location method in the steel structure building health monitoring system, analyzes the defects of the low node location accuracy of the traditional DV-Hop algorithm, and makes three targeted improvements. Firstly, four communication radiuses are adopted to refine the hops between nodes. Then, the weighted average hop distance is corrected by weighted processing. Finally, the improved sparrow search algorithm is used to locate the fault point. Taking the steel structure roof of the Terracotta Warriors and Horses Exhibition Hall No. 1 of Qin Shihuang Mausoleum as the simulation object, the MADIS_GEN software is used for modeling and dimensionality reduction. The improved ISSADV-Hop is compared with the traditional DV-Hop, IPSODV-Hop, and IGWODV-Hop algorithms in the fault point location of steel structure buildings. The results show that the normalized location errors are reduced by 19. 64% , 14. 87% and 8. 96% respectively, which verifies that the design scheme in this paper can effectively improve the positioning accuracy and is more suitable for fault point positioning in the steel structure health monitoring system.