Microgrids experience larger frequency deviations compared to bulk power systems for the same disturbance. Energy storage systems (ESSs) can potentially provide fast frequency support in such microgrids to limit frequency deviation within acceptable limits. One of the effective control approaches to achieve fast-frequency support in ESSs is a model predictive control (MPC)-based approach. Traditionally, MPC is known to use higher computational costs compared to other conventional controllers. In this paper, an MPC-based fast-frequency support mechanism is developed for an ESS and implemented on a real-time digital simulator to provide fast-frequency support in a microgrid model based in Cordova, Alaska. Results show that the computation time of MPC for frequency support is shorter than the simulation time step, justifying real-time applicability. The techniques presented in this paper can be generalized to develop novel MPC-based control approaches for ESSs and analyze their performance through real-time digital simulation techniques before deployment.