Renewable sources are important resources to be added to electrical energy systems in favor of decentralized generation. Discussions about environmental effects drive the studies in search of alternatives that make it possible to minimize the emission of polluting gases and diversify the electrical matrix. In this context, distributed generation (DG) of a renewable nature has been increasingly present, changing the classic structure of the system and giving greater engagement to the consumer. Therefore, it is essential to evaluate the performance of these new distribution grids in meeting the power demand, in order to establish appropriate standards and monitor them through the regulatory agencies. There is a diversity of methods for evaluating the performance of these networks, mainly through reliability concepts, to deal with equipment failures and the resulting effects. This thesis aims to evaluate reliability indices of distribution systems in the presence of DG. To this end, three techniques based on Monte Carlo simulation (MCS) are presented. A classic one, called sequential MCS, aims at modeling the chronological nature of the problem as well as the uncertainties arising from the intermittency of the DG sources. The second one, based on a quasi-sequential MCS, is characterized by its simplicity and ability to maintain the flexibility of sequential MCS, but with better performance in terms of precision and processing time. Finally, an MCS based on the concepts of chronological system state transition, also provides precision and flexibility to the reliability assessment. New test functions are proposed to achieve this goal. Two systems with DG insertion are used to evaluate the performance of the proposed concepts and techniques: IEEE RBTS, a standard academic network widely used in the reliability area; and another real network. The exhaustive discussion of the results confirms that both proposals meet the established objectives.