<p>The large-scale integration of renewable energy presents significant challenges to the control of modular multilevel converters (MMCs) in high-voltage direct current (HVDC) systems, which must contend with parameter perturbations, grid disturbances exacerbated by the intermittency of renewable energy sources, and nonlinearities. Traditional integer-order sliding mode control (SMC) suffers from chattering and requires prior knowledge of disturbance bounds, limiting its practical application. To address these issues, a novel fractional-order robust adaptive SMC (FO-RASMC) method is proposed in this paper. The proposed strategy integrates a fractional-order operator into the sliding surface for enhancing dynamic flexibility and reducing chattering, and introduces an additional switching gain to guarantee strict system convergence. Concurrently, an adaptive control law is constructed to estimate the unknown disturbance bounds in real time without prior knowledge, significantly improving control accuracy and dynamic performance while ensuring system robustness. The asymptotic stability of the proposed method is proved by Lyapunov theorem, and the dynamic convergence process and convergence time of the closed-loop system are quantitatively analyzed. Comprehensive simulations and controller hardware-in-the-loop (CHIL) experiments validate that, compared to conventional PI and SMC strategies, the proposed FO-RASMC exhibits superior dynamic performance, a wider stable operating range, excellent circulating current suppression, and reduced power oscillations under various scenarios. The CHIL results, obtained from a DSP-based real-time platform, further confirm the practical feasibility and real-time implementation capability of the method for MMC-HVDC applications, establishing it as a promising solution for future power systems with a high penetration of renewable energy.</p>