Smart building technologies are being developed in order to optimize energy usage in residential and commercial facilities electrical systems. One aspect of computer network based control systems is the intrinsic time-delays present in communicating feedback signals. This creates reduced stability margins in variable speed drives used for heating, cooling, ventilation and refrigeration systems. This research presents results for developing a model-predictive control method to compensate for the degraded performance associated with network time-delays. This is applied to a switched reluctance motor drive system. Improved stability margins are demonstrated for realistic time-delays encountered in wireless smart building systems.