Recently, more attention has been paid to distribution networks in power industry. They technological infrastructure is improved and the need for modern programs and methods is growing. They should be aimed at improving approaches of distribution network control. In the near future, necessity in state estimation distribution networks can arise because smart grids technologies are introduced in this area and gradually distribution networks are transformed into more complex structures. The most accepted state estimation approaches for any types of distribution networks are considered in the paper. In view of the differences between distribution networks and transmission grids, the usage of power system three-phase model is proposed for state estimation. A state vector in this task consists of phase parameters. The benefits and drawbacks analysis of three-phase state estimation approaches based on the weighted least squares method is presented. In this method a system state vector consists of currents or voltages written in polar or rectangular forms. The strengths and weaknesses of alternative state estimation methods applying are analyzed. Current distribution network problems are listed in the paper. They can be solved by making use of state estimation algorithms and other methods based on them. Furthermore, different approaches for network observability analysis of three-phase model, topology errors processing and bad data identification in a measurement vector are discussed.