Wide-Area Measurement and Control Systems (WAMCSs) will have a significant impact on a smart power grid implementation to enhance the power system stability. The dynamic performance of such WAMCSs can significantly deteriorate in presence of data dropout in the remote feedback signals from Phasor Measurement Units. Consideration of such an issue is important for any networked control system for ensuring efficient and reliable operation. To that end, in this paper, a new multi-input multi-output (MIMO) Observer-driven Reduced Copy (ORC) i.e. MIMO-ORC architecture is proposed for wide-area damping control using multiple doubly fed induction generator-based wind farms to mitigate the issue of packet dropout with multiple feedback signals. In this context, the concepts of cyber-physical self-coupling and cross-coupling are introduced and their impact on deterioration of closed-loop performance with data dropout is quantified through an analytical derivation. A framework for stability analysis of MIMO-ORC architecture is also presented. Finally, time-domain simulations show superiority of the proposed approach over its single-input single-output (SISO) ORC counterpart in a WAMCS for power grid.