Connected Electric vehicles (CEVs), as promising core factors in smart cities, take a substantial role in advancing the quality of core transportation services in Intelligent Transportation Systems (ITSs). The application of CEVs triggers the transportability and sustainability of smart cities, especially for ride-sharing and dynamic scheduling operations. Additionally, the combined optimization of the application of these schemes can further contribute to the potential benefits in smart cities. In this paper, an online CEV system for joint ride-sharing and dynamic V2G scheduling is proposed. Specifically, we formulate the joint problem as a mixed-integer quadratic programming (MIQP) problem. To deal with the forecast uncertainties, we thereby formulate an online scheduling problem, which also accounts for the communication effect to the real-time interactions. In the meantime, a bi-level system algorithm is devised to coordinate the CEV operations through the Benders decomposition. The simulation results show that our proposed model guarantees the effectiveness of the provision of both the ride-sharing and V2G regulation services. Through efficient CEV routing strategies, the active grid power fluctuations can be well alleviated by the provisioning of the V2G regulation service. Furthermore, these promising results contribute to large economic benefits in entire system operations.