Localization is a fundamental and essential problem for wireless sensor networks (WSNs). Many localization algorithms have been proposed. Making use of the unused or underutilized information which is able to be obtained in present hardware is a reasonable way to enhance the performance of the existing localization algorithms. In this paper, a local information-based range-free algorithm is proposed. We use the Received Signal Strength (RSS), which is unused in localinformation-based localization algorithms before, to compare the relative distances of two receiver nodes to the same sender node. According to the relative distance and considering the radio irregularity, we classify the anchor nodes into several kinds and choose some of them to localize the location-unaware node. In this way, the proposed algorithm is robust against radio irregularity. Moreover, we scan some selected grids, not all the grids, to estimate the location of unknowns, which decreases computation cost compared with an existing grid-scan algorithm. Analysis and the experiment results indicate that the proposed algorithm improves the localization accuracy and can localize more location-unaware nodes.