Abstract: The China national climate observation network composed of 2416 stations are taken as the reference station network, and correlations of surface air temperature series precipitation series between the reference station network and analyzed station network are used to investigate the monitoring ability for surface climate change. The relationships between the monitoring requirements and the station number are analyzed for 2.5°by 2.5°, 2.5°by 3.5°, 5.0°by 5.0° latitude/longitude grid cells. Results show that the needed amount of stations decreases rapidly with the lowing of ability (standard) required for monitoring. Taking 2.5° by 2.5° grid cell for example, the needed amount of stations decreases from 750 to 377 when the monitoring requirements for temperature change from 0.99 to 0.95, and the needed amount of stations decreases from 670 to 417 when the monitoring requirements for precipitation change from 0.95 and 0.90. With the increase of grid cells, fewer stations are needed for the same monitoring requirement. In areas of rugged terrain, undulating terrain and transition zone from mountains to the plains and plateaus, higher station densities are generally needed. The areas where need high-density stations are mainly located in the mountainous regions, including the Sichuan Basin, the Wuyi Mountains, the Loess Plateau, the Qinling Mountains, the Southeast Hills, and the eastern mountainous areas of Northeast China. The more complex the terrain is, the lower the spatial consistency of climate change is, and the more stations are required. In general, monitoring for precipitation change requires higher density of stations. For 5° by 5° grid, however, the difference of station densities needed between temperature and precipitation is relatively small. The cause needs further study. The distributions of stations at the central and the west parts of Tibetan Plateau and Taklimakan Desert are not enough due to the harsh natural environment. Further research is needed in these regions, favorably in combination with other methods. In addition, the current stations are often located in the valleys and foothills rather than those on the slopes or on the peaks in mountainous regions, which has lowered the representativeness and enlarged the uncertainties.