Abstract-In the Internet of Thing, data is almost collected, aggregated and analyzed without human intervention by machineto-machine communications resulting in raising serious challenges on access control. Particularly in Smart City ecosystems in which multi-modal data comes from heterogeneous sources, data owners cannot imagine how their data is used to extract sensitive information. Thus, there is a critical need for novel access control methods that minimize privacy risks while increase ability of personalized access control. Our solution is to build a trust-based usage control mechanism called TUCON that enables stakeholders to set access control policies based on their trust relationships with data consumers. In this study, we introduce two novel paradigms integrated in the Smart City shared platform: a Trust Service Platform and a Data Usage Control, then bring them together to establish the new mechanism. The conceptual model, the architecture, the formalization, and the practical development of TUCON is described in detail. We also show the roles and the interactions of TUCON components in the Smart City platform. Our contributions lie in a new trust model with a trust computation procedure based on semantic web technologies, a novel trust-based usage control conceptual model including a formalization, a practical expression and an architecture for Smart City systems. We believe this study provides better understanding on both trust and usage control in the Internet of Things and opens several important research directions in the future.