This work implements a MIMO (Multiple Input Multiple Output) MRAFC (Model Reference Adaptive Fuzzy Controller) to control the angular positions of an articulated 3DOF (3 Degrees of Freedom) robot arm. The structure of the MRAFC is represented by the Takagy-Sugeno fuzzy model. In this work, such a structure is enhanced via experimentation in order to fulfil the established design specification. The adaptive character of the MRAFC is necessary to overcome the parameter uncertainty of the modelling. Part of the design of the MRAFC requires the linearized state-space representation of the MIMO nonlinear model of the robot arm. Experimental results demonstrate that the implemented MRAFC is able to achieve the following desired design specifications for each degree of freedom: A settling time less than 2 s,a steady-state error close to null, and a P.O. (Percent Overshoot) around 2 %.