In smart city, which integrates the Information and Communication Technologies (ICT) including cloud computing and internet of things, all kinds of data collected by smart devices help to make everything intelligent. Through outsourcing encrypted data to cloud server, data owners can reduce the high storage and computational burden on resources-constrained smart devices. To further avoid the semi-honest-but-curious cloud server returning a fraction of false search results, this paper proposes a verifiable attribute-based keyword search scheme, thereby allowing users to check the correctness of the search results and achieving the fine-grained access control. Besides, our scheme considers the access privileges of the same data according to attribute-based priority tree. The formal security analysis proves that our scheme is secure against the chosen-keyword attacks in the generic bilinear group model, and the experimental simulations using a real-world dataset demonstrate its efficiency and feasibility in practice. © 2017 Elsevier Ltd