In this paper we present the multivariable controller design of a Lego Mindstorms NXT robotic arm for educational purposes. System identification was performed to estimate the parameters of a coupled plant. Uncertainties and Bode plots of the coupled system were evaluated to determine that the system could be decoupled at the DC gain value. A Proportional-Integral-Derivative (PID) controller was designed for the system after it was decoupled. These steps conform one of the basic techniques to design the controller for a multivariable system. Therefore, it is a valuable model to learn multivariable controller design for undergraduate students. This provides a hands-on approach experience in controller design where some aspects of advanced controller theory are blended with implementation details.