The presence of distributed generators (DGs) based on renewable energy is a fact in the electrical grid. However, DGs based on renewable resources such as photovoltaic panels and storage systems lack inertia, which is used by synchronous generators to compensate oscillations in the electrical grid. Thus, virtual inertia is introduced via a virtual synchronous generator (VSG) scheme. Although VSG is widely used, its robustness cannot be ensured since it employs proportional-integral-derivative-type (PID-type) controllers, which are sensitive to parameters variations. Furthermore, PID-type controllers are designed assuming balance conditions and the negative sequence components produced during unbalanced conditions are not considered. This paper proposes a robust VSG topology working under unbalanced conditions. A sliding mode control (SMC) algorithm named super-twisting (ST) is integrated into the control loop of VSG providing insensitivity to matched disturbances/uncertainties and finite-time convergence. Since the design of the ST algorithm considers the presence of negative-sequence components, it is not necessary a modification of the control loop before, during, or after the fault. The method to compute the ST control gains and the stability test using Lyapunov are provided. The VSG scheme is tested via simulations, where voltage sags are applied to generate the unbalanced conditions. Q r reactive power reference u st super-twisting algorithm α, λ super-twisting gains