Boilers are closed pressurized vessels which are used for heating water or generating water steam through combustion (combination of air fuel). In many commercial and industrial systems, boilers are used to heat the environment or provide the required process heat. Therefore, reducing their fuel consumption even slightly, can play an important role in reducing energy consumption and pollution. Nowadays, there are many ways to increase boiler efficiency. Closed-loop control of boiler is the most cost-effective and available method for increasing boiler efficiency and reducing pollution. In this research, a manually operated lab-scale boiler is setup for automatic control and monitoring. Values of boiler parameters such as air flow, fuel flow, outlet water temperature, inlet water temperature and exhaust temperature are measured by using image process algorithms and installed sensors and are sent to computer. Based on these measurements, values of efficiency and fuel–air equivalence ratio are calculated online. Some experiments are carried out on the boiler to obtain the operating point of the system, which is the desired fuel–air equivalence ratio in which efficiency is high and pollution is low. According to the conducted experiments it is determined that the boiler should work at a fuel–air equivalence ratio of 0.87. The control strategy implemented in this study is that at any moment the amount of fuel flow is determined, by controlling the equivalence ratio with reference 0.87. The desired air flow rate is calculated at any moment and it is given to the controller as a set point. Thus, by controlling the air flow, the fuel–air equivalence ratio remains in the desired value. The actual results of the implementation of this algorithm confirm its proper operation. Measurement of flows are delayed by 25 seconds due to the delay caused by image processing. For improving operation of the system, an inferential controller is added to the control algorithm to compensate the delay. This controller uses model of air actuator for computing air flow’s value when it is not available. The actual results of this algorithm are also presented in this paper. Key Words : Combustion Control, Lab-scaled Boiler, Image Processing, System Identification, Inferential Controller