For comparison purposes, this work implements two decentralized control systems to control simultaneously the water level and the temperature in a tank. The first control system employs MBPC (Model Based Predictive Control) controllers, while the second uses PID (Proportional Integral Derivative) controllers. Experimental results obtained in this work demonstrate that the performance using a MBPC controller is better than using a PID controller only in the level process. That is, the settling times and the steady-state errors were diminished. But in the case of the temperature process, the MBPC controller only diminishes the settling time.