Abstarct: State estimation is defined as attribution of determinated amount to an unknown state variable on a specific criteria that is presented through carried on measurements. Usually the number of measured quantities are more or less than state variable. So, the real value of state variables are estimated through state estimation. Even the measured quantities can be corrected via state estimation. The distribution networks consist of radial, short and un-transposed feeders with high R/X ratios and with single- phase for residential customers or three phase for both commercial and industrial customers. Therefore, the distribution networks are unbalanced in general. Furthermore, even if the distribution network is balanced, the asymmetrical faults additionally cause imbalanced conditions. So, it is suggested to use a model in network which can estimate system in these imbalanced conditions. For electrical system especially transmission level, single phase state estimation is common but in distribution network due to imballancity, single phase state estimation has significant error. In this study we are looking for improved method for three phase state estimation through investigating phases in distribution network. In this case, an accurate and reliable model in distribution system calculations is achived. Then, proposed model is demonstrated on modified 13 bus IEEE test system and evaluated the rate of its validity. Finally using an index and change in input quantities, the robustness of algorithm has been evaluated.