In this article, we propose an optimal decentralized robust $H_\infty $ control system for the voltage control problem of autonomous uncertain dc microgrids consisting of multiple distributed generations (DGs) with general topology. The one-degree-of-freedom (DoF) structure of the developed control system guarantees the robust performance of the system as well as its robust stability against various sources of uncertainty, such as plug-and-play (PnP) operation of the DGs, topology changes, large load perturbations, different subsequent system changes, and the presence of constant power loads (CPLs). To that end, the uncertain dc microgrid is modeled as a Linear Time-Invariant (LTI) polytopic state-space system. Then, a state-feedback control technique based on linear matrix inequality (LMI) with linearly parameter-dependent (LPD) Lyapunov matrices is implemented on the uncertain dc microgrids. The proposed controller does not require any sort of communications, and the robust performance property of the controller eliminates the need for prefilter design. The efficiency of the proposed controller is analyzed by simulating different case studies through SimPowerSystems Toolbox in MATLAB. Finally, the performance of the proposed control method is validated via experimental studies.