This thesis tackles the high energy use that limits wider adoption of controlled environment agriculture (CEA), even though CEA can deliver reliable, high quality crops regardless of weather. It proposes a cyber physical biological systems approach that brings together physics based modelling, machine learning, and robust optimal control to manage the indoor climate and resources in semi closed greenhouses and plant factories.First, the thesis develops an integrated model that links weather, climate control actions (temperature and humidity), CO2 enrichment, irrigation, and fertilization to the indoor environment and crop states. A simplified, linear version of this nonlinear model is also built to keep computation fast while preserving accuracy. On top of this, a robust model predictive control framework is designed to improve the use of renewable energy and keep growing conditions within target ranges. The framework learns uncertainty from past forecast errors so it can make strong decisions without being overly cautious. The thesis also introduces physics informed deep learning to create digital twins—high fidelity virtual models—of the indoor climate and crop growth. These twins support prediction and control in plant factories under different daylight conditions. Finally, a hybrid reinforcement learning and robust MPC method learns changing uncertainty patterns directly from weather data, removing the need for a fixed robust formulation and speeding up decision making. The methods are tested in realistic simulations for tomato production across different facility types and climates. The results show that the proposed approach maintains target environments, improves energy and resource efficiency, lowers operating costs, increases productivity, and reduces violations of climate and growth limits when compared with conventional control strategies. Overall, the thesis delivers practical, data efficient control tools that connect plant biology with sensing and control, enabling more sustainable, lower carbon CEA and supporting broader use of renewable energy in food production.