Abstract: Indoor landmarks play an important role in the location-based service (LBS). On account of outdoor landmark extraction methods can not be fully applicable to complex indoor environment, we propose a suitable indoor landmark saliency quantitative evaluation model to extract indoor landmarks. In this paper, large shopping malls are regarded as research site, after analyzing the factors influencing the saliency of POI objects from perceptual, cognitive and spatial structure, a saliency measure model composed of these factors is constructed. Finally, we compute the significances of the POIs selected from the interior of Chicony shopping mall of Wuhan city. In this experiment, several level of landmarks are extracted to reflect different size of spatial knowledge in indoor spaces. Extracted indoor landmarks can be important symbols in indoor intelligent navigation system and as key clues in a large shopping malls wayfinding, multi-granularity route directions.