Abstract: In the process of indoor location based on RSS fingerprint, the quality of the obtained similar point set is a key factor for a successful position. And the locating point’s RSS is an important reason which affects the quality of the similar point set. By analyzing the distribution of D-RSS, the concept of RSS’s typicality was proposed firstly, and an indoor localization algorithm based on typicality judgment of RSS was also presented. According to the principle that the RSS values and the effective similar sample points, a typicality discrimination method for RSS values and a self-adapting K value were presented. Confirmed by the experiments, the algorithm not only can find the effective similarity sample points completely, but also can eliminate the non-substantive similarities points, and then can adapt to the different scenes, then have the higher positioning accuracy.