This thesis focuses on developing current controllers for industrial-grade DCDC converters used for voltage regulation in direct-current (DC) microgrids. These converters are controlled through droop control, which consists of two main loops: an inner current regulator (the focus of this thesis) and an outer voltage regulator based on the DC droop curve. This curve defines the relationship between the DC link voltage and the current drawn from the converter, with high-performance droop controllers allowing flexibility in shaping this curve. Crucially, the curve’s shape is limited not only by the hardware components but also by the inner current control loop. Therefore, this work investigates current controllers using optimal control techniques to achieve reliable current regulation and enhance the droop controller’s performance. Initially, a dynamic model of the converter is obtained, followed by the development of an estimator to compensate for measurement delays. Various current controllers are devised using the linear quadratic regulator (LQR), linear model predictive control (MPC), and nonlinear MPC (NMPC). Simulations demonstrate significant performance improvements over the existing Ackermann’s pole-placement method. Given the device’s microsecond operation range and limited computational resources, direct deployment of MPC-based controllers is infeasible. To address this, an approximation of the linear MPC controller is derived using imitation learning. This approximated control law is a polynomial function with a small memory footprint and fast evaluation time, making it suitable for embedded devices. The LQR and the approximated MPC controllers are implemented on real hardware and assessed on a test bench, confirming their superiority over the existing Ackermann controller.