Electric vehicles’ market penetration has been rising due to new technological developments and awareness of the climate change threat. Instead of being a burden for today’s electricity production, such as increasing peak demands, they can also have a positive impact when charged smartly, coupled with clean solar electricity (PV). Smart charging models have to be able to account for the uncertainty of production and consumption forecast, the driver’s satisfaction, and other objectives such as operational cost, PV self-consumption, and peak shaving. The goal of this thesis is to build a smart charging model capable of forecasting uncertainty, include it in the optimization problem and implement real-time solution adjustment through dynamic control. The results show that combining Artificial Intelligence, stochastic optimization, and Model Predictive Control in a smart charging model leads to a decrease in the cost of operation, an increase in PV-self consumption, and reduces peak demand to the grid.