Worker selection is very crucial for crowd-sensing to ensure high data quality. Existing approaches have two limitations. First, they only take specific factors into account for their motivating application scenarios, but do not provide general models in support of crowd-sensing at large. Second, they select workers only in terms of the requirements defined by the task creator without considering other worker-required factors. To overcome abovementioned limitations, this paper proposes a novel worker selection framework for crowd sensing. Compared to existing work, it mainly has following two characteristics. (1) Multi-scenario. Instead of defining specific factors, we propose a core ontology model to semantically express general factors, based on which task creators can build their own task-specific models efficiently. (2) Multi-view. We propose a two-phase process to select workers by considering factors both from the task creator and worker. We evaluate the effectiveness of the worker selection process by using a questionnaire-generated dataset. Results show that our approach outperforms the baseline method. © 2015 IEEE.