The paper presents LDM-SG (Latvian Dynamic Model for Distribution network reliability) simulation software to be used for the technically-economic analysis of power supply reliability for urban distributed medium voltage (6, 10, 20 kV) networks with distributed generators. The following aspects are described: 1) software structure, 2) main module functions, simulation facilities and algorithm principles. The paper considers in more detail the methods for power supply reliability assessment and simulation as applied to local power stations. The main modules of the software are: 1. Options. A module for forming options and comparing their technically-economic criteria. 2. Electrical calculations. A module for computing the main electric parameters in normal and post-emergency operating conditions as well as the reliability criteria in 10 (20) kV and 110 (330) networks and the reliability criteria for switchgears at distribution points. 3. Analysis (decision making). A module for reviewing different forecasts (increase in the load, interest percentage, etc.) and performing the sensitivity and risk analysis. The software modules can be used separately and in different combinations and sequences. The model architecture makes it possible to solve various complicated simulation problems thus providing the opportunity to perform dynamic optimization and sustainable formation of distribution networks. The LDM-SG software allows for performing dynamic optimization when analysing different options, taking into consideration technically-economic parameters of reliability. The LDM-SG software enables the following: • entering the entire scheme of a network; • formation of development options, electrical computations for the normal and post-emergency operating regimes; • comparison of the economic effectiveness for different options; • risk analysis under information uncertainty. As far as the authors know, no software exists with such possibilities for reliability analysis. In their work, the authors have applied the probability theory methods widely acknowledged all over the world.