Sensitivity analysis of the biogeophysical model called CH4MODwetland showed that environmental driving factors such as temperature and standing water depth are the most sensitive contributors to CH4 fluxes from natural wetlands. The response of CH4 fluxes to standing water depth is more sensitive in seasonal flooded wetland than in continuously flooded wetland. The sensitive significance of the model to input parameters increased in the order of the maximum aboveground biomass, the proportion of below-ground biomass to total production biomass, the vegetation index, the required accumulated temperature for reaching maximum biomass, and soil input parameters. The empirical equations of water table are used to support the CH4MODwetland in order to simulate the variations of CH4 fluxes from the 1950s to the 2000s in Sanjiang Plain and from the 1960s to the 2000s in Zoigê Plateau. The interannual change in CH4 fluxes is mainly affected by precipitation, while the climatic warming enhanced the CH4 fluxes from wetlands during similar precipitation years. In the Sanjiang Plain, the simulated CH4 fluxes increased by 9.5% and 8.3% from the 1980s to the 2000s compared with those from the 1950s to the 1970s in the marshland dominated with Carex lasiocarpa and Deyeuxia angustifoli,respectively. In the Zoigê Plateau, the simulated CH4 fluxes increased by 6.0% and 5.5% from the 1990s to the 2000s compared with those from the 1960s to the 1970s in the peatland dominated with Carex meyeriana and Carex muliensis, respectively. The results of this study may be used to project the effects of future climate change on CH4 emissions from Chinese wetlands.