The service interaction provided by objects in IoT networks enables the creation of advanced services to answer application requests. However, the growing number of objects into the IoT network, besides its ad hoc structure, are disturbing some functionalities, such as service discovery. Therefore, when searching for services, the navigability is impaired because the system needs to sweep a great quantity of objects without a previous organization. Social Internet of Things (SIoT) emerged as an alternative to solve several problems faced by IoT through the concept of social networks. In SIoT each object has its own social profile, which contains its characteristics and information, and are organize by relationships. Thus, this research propose a solution for service discovery in a SIoT network. This solution uses the relationships between objects to improve the discovery scalability and considers their social profiles to meet more satisfactorily the requisitions. Simulated results demonstrates the solution performance to answer service requisitions in an urban SIoT network.