With ever increasing penetration of renewable distributed generation (DG) in distribution systems, power restoration of remote distribution feeders under emergency conditions tends to be carried out with the support of renewable DG units. The available power from the renewable DG units ensures restoration of more number of affected customers, thus, improving overall system reliability. In this paper, a probabilistic based analytical method is developed to assess system reliability in terms of system average interruption duration index and system average interruption frequency index for distribution feeders containing dispatchable and nondispatchable renewable DG units. The proposed method has been developed by implementing DG side restoration with comprehensive technical considerations, including possible failures of DG units, time-dependent patterns of load demand and DG power output, and single-stage and two-stage restoration. The proposed analytical method is validated by comparing with the Monte Carlo simulation and results are presented.