A new, improved model for planning of low–voltage distribution networks is presented in this dissertation. The model uses hierarchical and stochastic optimization methods, as well as modern programs for drawing and load flow calculations. From the group of hierarchical methods, fuzzy clustering method and analytic hierarchy process are used in the model. Usage of these methods gives the best spatial solution of low–voltage distribution network. From the group of stochastic methods, simulated annealing method is used in the model. Usage of this method gives the optimal switching condition of the best spatial solution. For drawing the spatial solutions of distribution network AutoCAD is used in the model, while for load flow calculations Matlab–Matpower is used. In the first part of the dissertation a short theoretical assumptions about distribution networks are given, as well as mathematical interpretation of optimization process and used optimization methods. In the second part, a new, improved planning model, the results of model application to the specific geographical area or group of customers, as well as the comparison of results with the existing distribution network are presented. From the results obtained by the usage of proposed model it can be seen that a series of spatial solutions is obtained, and the designer can select the most appropriate. Also, the model provides the possibility of obtaining the optimal switching condition of selected spatial solution. Comparable results obtained by model implementation and existing distribution network show significant improvement of network parameters. That is a practical confirmation of a justification for the proposed planning model application. Its application is particularly evident in unplanned built distribution systems, where there is a large number of similar networks, with distinct problems in exploitation, so their resolution can give a significant financial savings. The ability of the model to detect problems in the existing networks and verification of proposed solutions is an additional advantage. One of the advantages of the model is the ability for usage the other optimization methods as well, that are more acceptable for other designers, depending on their programming skills, but the same steps in the implemented procedure have to be used. In the dissertation is presented an example of low–voltage distribution network, but the model can be successfully used for designing the medium–voltage distribution networks, which further expands its scope of applicability. Considering the simplicity of the procedure that is necessary to be conducted for experienced engineers and planners of distribution networks the model can serve as a very useful tool and has its place in the practical application.