This work proposed a single base station positioning design in 5G networks, which can jointly estimate user equipment (UE) distance (time of arrival, ToA) and direction (angle of arrival, AoA) by utilizing the wideband 5G signal and vector antenna (VA). A statistics-based Expectation-maximization (EM) algorithm and a subspace-spaced algorithm are adopted to estimate the UE position in this work. The simulation results show that the proposed method can accurately estimate UE position by using the uplink sounding reference signals (SRS) in a Line-of-Sight (LoS) scenario where the tapped delay line D (TDL-D) channel model is used to construct delay and attenuation profiles for multiple paths. Also, the impact of the strong reflections on angle estimation and polarization accuracy is studied in the subspace algorithm. This work proves that the VA is able to provide high accuracy 3D UE positioning in 5G networks without the requirement of multiple cells or multiple antennas. However, the performance of the VA antenna is limited by the coarse angle resolution, which needs to be resolved by VA composed antenna array in future works.