The satellite signals in city street are easily blocked and the accuracy of Global Navigation Satellite System (GNSS) direct positioning is poor. This paper proposes the GNSS precise positioning algorithm of Unscented Kalman Filter (UKF). Firstly, taking into account that the measurement of the pseudorange and carrier phase observations is seriously affected by the Non line of sight (NLOS) propagation of the GNSS signal during transmission, the Support Vector Machine (SVM) classification algorithm is used to classify the signal so as to obtain ideal GNSS signal. The relationship between the neighboring pedestrians and the target pedestrian is then established through mathematical modeling. Then, the satellite positioning data of the neighboring pedestrians are used as the observation data to correct the positioning error of the target pedestrian by Unscented Kalman Filter (UKF), thus achieving the goal of improving the positioning accuracy of target pedestrian.