The main purpose of IoT is to deliver reliable, high quality services and innovative solutions by transforming the captured data into meaningful information, and thus improving user's daily life. In this regard, it is in the interest of the community to encourage entities within IoT environments to share their data, and therefore serve public interest and contribute to the innovation and technological progress. Meanwhile, the distributed nature of IoT networks and the diversity of its actors lead to the recognition of security and data sharing management as one of the major challenges of the IoT domain. For instance, due to insufficient governance of the shared data within IoT environments, data provider retains little to no control over his assets once he has agreed to share them. Furthermore, data consumers are not able to trace the source of the available resource nor its history processing to assess its quality. All this creates a digital environment that is certainly functional but lacks mutual trust between its actors, which can prevent the domain's full potential to be exploited, and therefore disrupt the implemented services. In our work, we propose an approach to improve data sharing management using three main elements: semantic modeling, usage control policies, and data provenance.