In a previous paper (Benchoubane and Stoten 1992) we noted that the minimal control synthesis (MCS) algorithm, when applied to the control of plant with rapidly varying disturbances, could lead to stable, but only bounded, error dynamics. The purpose of this paper is to show that, with some modification, MCS can elicit global asymptotically stable error dynamics, even in the presence of rapidly varying disturbances. We label the modified algorithm as the extended minimal control synthesis (EMCS) algorithm. Disturbances include plant parameter variations, external inputs and plant nonlinearities. In common with the general philosophy behind MCS, we assume that all nominal plant parameters and disturbance parameters are unknown. Two examples of EMCS implementations—one for a SISO plant and the other for a MIMO plant —are included in the paper to demonstrate the effectiveness, and practical aspects, of this new algorithm.