Abstract：GPS data from continuously operating stations both in Nepal and in the southern Qinghai-Tibet plateau are processed and three-dimensional transient postseismic displacements are calculated during the first year after the earthquake. Significant postseismic displacements concentrate in northern Nepal, especially along the national boundaries between Nepal and China. The amplitudes and azimuths of the horizontal postseismic motions exhibit a pattern similar to that of the coseismic offsets. Almost all stations in Nepal record postseismic uplift. Kinematic finite fault afterslip models are constrained using GPS measurements. The preferred afterslip model shows that the afterslip occurred in the down-dip portion of the coseismic rupture with an average slip-depth of 20 km. The maximum depth of the raw afterslip model extends to 30 km. However, subtle afterslip is detected in the shallow portion of the fault. The moment release is estimated to be 8.8×1019 Nm, equivalent to a magnitude of Mw7.3 and approximately 12% of coseismic moment. Then, the residual afterslip model is inverted by using the residual data set between the raw GPS data and viscoelastic models, which is based on lateral rheological structure. The residual afterslip model shows narrower and shallower afterslip located in the downdip of the rupture with an average slip-depth of 16 km. The moment released by the pure afterslip model decreases to 5.7 ×1019 Nm, corresponding to Mw7.2. The afterslip-coseismic moment ratio drops to 8%. Our simple multiple mechanism model not only improves the data fit but also makes the residual afterslip distribution more physically reasonable. Our afterslip models show no slip within the shallow unbroken portion of the fault in the postseismic period, implying it is accumulating elastic strain and will be released through earthquake in the future.