An effective way to extend to the multi-input case the variable structure control philosophy is based on a set of m + 1 control vectors forming a simplex in R(m) and on the corresponding snitching of the controlled system from one to another of m + I different structures. Yet, some problems arise when uncertainties are present in certain matrices characterizing the controlled systems. In this paper, conditions are identified under which, even in the presence of uncertainty, the convergence to the sliding manifold is ensured via the application of a multi-input control strategy still based on a simpler of control vectors.