Rule elicitation remains the most crucial problem in the design of a fuzzy logic controller. This is even more difficult if the process is multivariable, in which case the number of rules increase exponentially with the number of variables. To overcome this, a different type of learning fuzzy control algorithm is presented. The control method has been called the model reference fuzzy adaptive control (MRFAC). This algorithm uses a reference model (which specifies the closed-loop process behaviour) to provide performance feedback for synthesising and modifying a fuzzy controller's rule-base. A multivariable process has been used to test the MRFAC system and the results are presented.