This chapter proposes a novel control method of integrating online learning with computational fluid dynamics (CFD) to control the temperature in a generic combustion process, and discussion comparing the method with a proportional–integral–derivative controller is included. The dynamic system consists of a metal block that exchanges heat with the environment. A heater, which is controlled by the online-learning controller, is situated inside the glass-enclosed system. The online-learning controller is trained using real-time data, which flows from the online-simulation module rather than the plant processes. CFD-based online-simulation technology is used to simulate the combustion process, which obtains the real-time data that the controller needs to implement in a control loop for the boiler-combustion process. In a real combustion process, if the load of a boiler is required to change frequently, that also means the fuel control has to change frequently to meet the load requirement.