The aim of this work is to maintain the interacting liquid level and temperature parameter at a desired level. This paper presents decoupling, linearization algorithm (Hirschorn's algorithm) and Kalman filter (KF) and Extended Kalman filter (EKF) for an approximated model of an interacting thermal nonlinear process. The nonlinear interacting system is converted into linear non interacting system using decoupling and linearization algorithm. KF and EKF are then designed to estimate the system parameters namely level and temperature for a non-interacting linear Multi Input Multi Output (MIMO) system. Performance of Extended Kalman filter was found to be better. The obtained estimated error for the plant using EKF is less when compared to KF.