In order to obtain the displacements of observation points,high-rate GPS data sets are utilized to compute the positions epoch-by-epoch with precise point positioning (PPP) techni-que. However,the precision of PPP technology is affected by systematic errors that are not currently modeled. This paper investigates high-rate GPS positioning time series and describes methods for weakening these effects. Spatial stacking and sidereal filtering methods are introduced to improve the PPP positioning results. Some high-rate GPS observation data at permanent sites from San Simeon earthquake in America and Wenchuan earthquake in China,respectively,are calculated and analyzed. Numerical results demonstrate that stacking method could efficiently eliminate spatial correlation error between stations. While sidereal filtering has positive effect on removing multipath error. Both methods could be used in reducing positioning noise involved in PPP,and make it easier to accurately detect weak earthquake signals.