In view of the shortcomings in localization accuracy of current multiple target localization algorithm, a method of multi-object localization based on compressive sensing (CS) and alternate iteration is proposed. Firstly, the measurement matrix of received signal strength (RSS) is expressed as the product of measurement matrix and sparse vector according to CS theory, which transforms multiple target localization problem to the reconstruction of sparse vector. Next, traditional localization algorithm based on CS is presented to obtain rough estimation of target positions. Finally, alternate iteration method is employed to further refine positions when localization results are not accurate. During the alternate iteration process, the diamond search is used to find the exact target locations. Simulation results show that the proposed algorithm overcomes the limitation of traditional compressive sensing localization techniques which can only locate targets in the center of grid, and improves the localization interference of the interaction between objects with high localization accuracy. The indoor operation and maintenance inspection area of a power company in Chongqing is chosen as an experimental site and the proposed method is applied to the actual inspection localization, and good results are achieved in the indoor localization.