Smart environments rely on sensor data to provide necessary business intelligence in order to support decision-making. An efficient decision support model in such a context requires that sensor data are provided correctly and timely. Given the dynamicity of sensor environment, the diversity of their features and of user requirements, finding appropriate sensors having the required capabilities or replacing faulty ones constitutes a challenging task. Efficiently describing and organizing sensors in smart environments is essential to deliver a rapid adaptation to errors and availability of data. In this paper, we present an approach for organizing and indexing sensor services based on their capabilities. We introduce a feature-oriented capability model that puts forward the functional aspects of carried actions and model them as resource description framework (RDF) properties rather than focusing on the change in the state of the world. Using this model for describing sensor capabilities, we apply Formal Concept Analysis for organizing and indexing sensor services. We have experimented and evaluated our approach in the Digital Enterprise Research Institute, which has been retrofitted with various sensors to monitor temperature, motion, light and consumption of power within a building.