Bidirectional DC–DC converters play a critical role in DC microgrids by enabling regulated voltage conversion ‎and bidirectional power flow between energy sources, storage systems, and loads. However, achieving fast and ‎robust voltage regulation remains challenging due to non-minimum phase dynamics, digital ‎implementation delays, and model uncertainty. Short-horizon finite-state voltage control strategies are introduced ‎as desired methods for controlling these converters in the literature. In this paper, a systematic comparison of ‎short-horizon finite-state control strategies—including single-horizon finite-set model predictive control, ‎compensated bang-bang control, and bang-bang control with doubled sampling frequency—is first presented. ‎Subsequently, a novel model-free bang-bang control strategy is proposed. The proposed approach ‎eliminates reliance on an explicit system model by generating desired inductor-current reference directly from ‎measured current signals, while preserving short-horizon operation and low computational complexity. The proposed method is then compared with existing approaches through comprehensive ‎simulation studies under non-minimum phase operation and digital implementation delays. The results demonstrate that the ‎proposed model-free strategy achieves transient response, settling time, and steady-state voltage ‎regulation comparable to the model-based methods, while eliminating dependence on converter parameter ‎accuracy. These findings highlight the effectiveness and practical suitability of the proposed approach for ‎digitally controlled bidirectional DC–DC converters in DC microgrid applications.‎