This paper presents two approaches for the fuzzy sliding mode control of uncertain and disturbed non-linear multi-input multi-output (MIMO) systems. In the first approach, the control law is deduced from a nominal model of the plant. An adaptive fuzzy system is introduced to substitute the switching control and, hence, to eliminate the chattering phenomenon. In the second approach, two other adaptive fuzzy systems are used to approximate and thereby to overcome the unavailability of the nominal model. The stability and the robustness are proven analytically in both cases, and the adaptation laws of the fuzzy systems are deduced from Lyapunov synthesis. A simulation example of a two-link robot is used for illustration.