This work presents a comparative study of PI and PID controllers applied to both linear and non-linear tank level processes. A single tank level control process is considered to represent a linear system, while a conical tank illustrates nonlinear behavior arising from variations in cross sectional area with height. The process dynamics are determined through a step response system identification method and approximated by an order of one included with dead time (FOPDT) model. Controller tuning is performed using the Ziegler-Nichols (ZN) method. Simulation results are obtained for different set-point changes at various operating conditions. For the linear case, both PI and PID controllers provide satisfactory tracking performance. However, in the nonlinear conical tank, the PID controller tends to destabilize the system under negative step changes, highlighting the challenges of nonlinear dynamics and the importance of appropriate controller selection. This study suggested an I-PD controller for controlling a nonlinear conical tank, which is tuned using Ziegler-Nichols method, the controller tracked -5% to -20% set-point changes accurately. Compared to a conventional PI/PID controller, the I-PD provided smoother and more stable control. The simulated results show that proposed I-PD controller is better for nonlinear processes.