For the problem that the observation environment is poor and the positioning accuracy is not high, pseudolites can be used for positioning. However, compared with navigation satellites, pseudolites are much closer to user receivers. Using algorithms such as EKF may cause relatively large linearization errors, but UKF algorithm can avoid linearization errors. Through theoretical analysis, it is found that the distance between the satellite and the receiver of the user and the coordinate error of the user receiver will affect the linearization error. In this paper, We used different receiver coordinate errors, and a set of short-baseline pseudo-satellite positioning experimental data was solved by UKF algorithm and EKF algorithm to verify the influencing factors of linearization error. The experimental results show that the linearization error is larger when the receiver coordinate error is larger or the satellite distance from the receiver is smaller. When the receiver coordinate error is relatively large, the convergence speed and overall accuracy of the UKF algorithm are better than EKF algorithm.