This thesis addresses Neural Network Based Predictive controller based control systems for coupled tank system. The first is a PID controller based control systems. PID controller tuning methods has been carried out. Neural Network based Predictive Controller and system Identification has been discussed where controllers are designed for single rectangular tank system, coupled rectangular tank system. Initially PID Controller is introduced for Single Rectangular Tank System. The controller is tuned by Manual Tuning Method and Ziegler-Nichols Tuning Method. For controlling these two systems Neural Network Based Predictive Controller is suggested. Block specification, system identification and Back propagation are also illuminated. The proposed technique shows the satisfactory result in terms of controlling the coupled tank system with fixed input and it is also executed in simulation and gives the satisfactory result. Finally process is reliable for nonlinear plant in simulation. At last, MATLAB coding produces different response curve for real time system which is also discussed in the paper.