This paper presents an application of a state estimation algorithm as a tool for re-synchronization of distributed generation in a distribution system. The estimation results can be used to operate a distributed generation in grid and off-grid mode. The estimated state variables are utilized as input data to control a Diesel generator while resynchronizing as well as for a grid mode loss predictor. The proposed state estimator, with branch currents as state variables, was solved by the weighted least squares method using singular value decomposition. To ensure the observability of the distribution system, a detailed numerical observability analysis was carried out with complex power measurements and unconventional branch current phasor measurements. In the algorithm, phasor measurement units and pseudo-measurements can be incorporated. Conditions that ensure system observability and redundancy increasing are derived when those measurements are included. The approach was tested in the IEEE 13 node test feeder, to which a distributed generation was connected for external network feeding. The algorithm shows stability, convergence, allowing the distributed generation re-synchronization to the distribution system. This feature could help improve self-healing capability in microgrids that use synchronous generators.