Aiming to explore the evolution of Global Positioning System(GPS) station's movement characteristics after the Nepal earthquake in southwestern Tibet, the daily position time series of 7 GPS stations located in southwestern Tibet spanning from 2013 to 2018 are used. The common-mode error (CME) of the regional GPS network is extracted using principal component analysis (PCA). The contribution rate of the first principal component (PC1) in the up component rises from 48.87% to 64.39%, and the amplitude of the up component of CME is enlarged after the earthquake. Maximum likelihood estimation (MLE) is used to analyze the impact of seismic event on the noise characteristics of GPS position time series. We mainly pay attention on white noise (WN) and power-law noise in the noise analysis. After the earthquake, random walk noise (RWN) with relatively large amplitude appears at XZAR, XZZB and XZZF stations. For the WN+FN (flicker noise)+RWN noise model, the white noise amplitude of most GPS position time series is decreased, while the amplitude of flicker noise plus random walk noise is increased. The average spectral index of GPS position time series in the north, east, and up component varies from -0.98, -1.07 and -0.98 to -1.27, -1.15 and -1.03, indicating that the percentage of white noise in the GPS time series is decreased, and the power-law noise dominate the total noise. The velocity field of some GPS stations in the horizontal component appears to shift southward, especially at XZAR, XZZB and XZZF, and the movement rate also is decreased accordingly. In terms of annual and semi-annual cycles, the amplitudes of seasonal signals 6 GPS stations is enlarged with different amplitudes, while only the station LHAZ is slightly decreased. The above results imply that the Nepal earthquake may has changed the movement characteristics of most GPS stations in southwestern Tibet, and then suggest that the earthquake affect the original crustal movement over the southwestern margin of the Tibetan Plateau.