The Internet of Things (IoT) perspective endeavours to unite all the physical objects or ‘things’ embedded with electronics, software, sensors, and network connectivity to allow more direct integrations between the physical world and cyber-based systems. At the same time, these networked devices and associated data communications can increase the energy demands exponentially and could seriously harm the environment because of their carbon footprints. Currently, most researchers and practitioners have dedicated their efforts to improving the energy efficiency of IoT systems. But most are falling into a box of focus within telecommunications and Internet domains themselves. This research targets to develop a novel, eco-friendly and sustainable data transmission approach to offload the data traffic from infrastructures to opportunistic and social networking by exploring the existing movements and spatial closeness relations among things. Objectively, to complement the traditional infrastructure-based data transmission, the new idea is to optimally piggyback the data on the moving physical objects/things for the data delivery to achieve energy reduction and ensure their Quality of Services (QoS) requirements. The thesis contribution is presented in three phases; in the first phase, the idea of similarity in the user’s mobility patterns is explored. These similarity patterns are found using the co-occurrence matrix. The identified patterns further motivate the second phase, where the research predicts encounters among mobile users. This enables the approach to investigate different mobility patterns and find common points of meeting among other users. The meeting points are then predicted using the random forest approach. After getting the accurate results of ‘prediction of meeting (encounter)’. The multiple-criteria decision model will decide based on delay and transfer the data through the Internet or Device-to-Device (D2D) in the third phase.