This paper presents is to develop and compare neural network and conventional based controllers for a boiler of steam power plant. Designs of two different controllers for pressure and temperature are presented for keeping the boiler working in normal condition and improve efficiency. These controllers consist of NARMA controller of ANN and a conventional proportional-integrator-derivative (PID) controller. These parameters are adjusted by built a model and implementation in MATLAB program according to the requisite of the steam power plant and the control objectives. The results show a neural network is best controlled and superior performances of power plant from PID controller artificial neural network and PID have been applied in Al–Dura power plant in Baghdad. Therefore, neural networks have been extensively utilized in many industrial applications.