Boiler drum level control in thermal power plants aims to balance the input water flow and the superheated steam in the output of the boiler. It is assumed to be the heart of the boiler control system including hundred parameters need to be observed and controlled. Currently, most thermal power plants are applying classical PID regulators to control the water level of the steam drum. The disadvantages of these controllers include the difficulty of parameter update ability, manual regulation and poor control performances. With the development of the modern control theory, it is effective to replace the conventional PID regulators with intelligent controllers. This paper proposes two smart control strategies applying fuzzy logic and neural network to stabilize the drum water level. An artificial neural network-based controller, namely NARMA-L2, is the novel contribution of this study, dedicating the better control performances in comparison with the conventional control scheme as well as the fuzzy logic based controller. Simulation results using MATLAB/Simulink package with real parameters of Pha Lai 2 conventional thermal power plant verified the efficiency of the proposed control strategy.