Reliability is an indispensable factor in power system design and operation and has a significant impact on grid safety and economy. Future power distribution systems are expected to be more sophisticated, owing to the increasing penetration of renewable resources and adoption of advanced information and communication technologies. Extant studies in this field tend to focus on the modeling and assessment of the reliability of future microgrid distribution systems, including distributed generation, without considering networked configuration and limited transfer capacity. In the work presented in this paper, a Markov model is implemented to perform a practical and accurate reliability evaluation of networked electric microgrids under load transfer restriction conditions. The Markov model is used to model the microgrid based on the connectivity between the source and the loads and to compute load and system reliability indices. Moreover, the distribution load flow (DLF) method is adopted when reclassifying the Markov model states based on the system's transfer capability during interruptions. The obtained results confirm that the proposed model is efficient and that the DLF provides a more accurate reliability analysis due to the computing of the voltage profile during the system outage restoration process. This model can also be used to optimally integrate distributed generators into the power system at proper locations and with proper capacities to enhance the system's reliability.