The various types of ecosystems and complex landforms found in the cold alpine area of southwestern China make this region ideal for researching regional responses to global climate change. Therefore, to evaluate the responses of regional carbon and water cycles to climate change; it is of great importance to investigate the response of water use efficiency (WUE) to the climate in this region. A process-based ecosystem model, Carbon Exchange between Vegetation, Soil, and the Atmosphere (CEVSA), was used to estimate temporal and spatial variations of WUE in the terrestrial ecosystems in the alpine area of southwestern China during 1954-2010. First, we ran the model using the average climate data from 1954 to 2010 until an ecological equilibrium was reached, then we conducted dynamic simulations with climate data at a time-step of 10 days during the same period. Moreover, the correlation coefficients between WUE and climate variables were calculated to analyze the relative effects of temperature and precipitation on variations of WUE. To achieve the results, various types of computer software were used, such as ANUSPLIN4.1, Fortran 95, Arcgis9.3, and SPSS18.0. The results showed that the average WUE in the studied region was 1.13 g C mm^-1 m^-2 during 1954-2010. The mean WUE of three main vegetation types included 1.35 g C mm^-1 m^-2 for herbaceous cover, 1.14 g C mm^-1 m^-2 for evergreen needle-leaf tree cover, and 0.99 g C mm^-1 m^-2 for evergreen broadleaf tree cover. In spatial distribution, significant positive correlations were found between the annual WUE and altitude (r=0.156, P < 0.05), and significant negative correlation was found between the annual WUE and annual mean temperature (r=-0.386, P < 0.01). Moreover, the annual mean WUE in the entire region showed a significantly decreasing trend at a rate of 0.006 g C mm^-1 m^-2 a^-1 (P < 0.01). Significant negative correlations were found between the annual mean WUE and annual mean temperature (r=-0.727, P < 0.01), and no significant correlations were found between the annual mean WUE and annual precipitation. The decrease in WUE resulting from an increase in evapotranspiration (ET) was more than that of net primary production (NPP) from the temperature increase during the study period. Furthermore, decreasing trends were highly significant in herbaceous cover at 1.37 × 10^-3 g C mm^-1 m^-2 a^-1, evergreen needle-leaf tree cover at 6.17 × 10^-3 g C mm^-1 m^-2 a^-1, and evergreen broadleaf tree cover at 1.03 × 10^-2 g C mm^-1 m^-2 a^-1 during the study period. The annual WUE showed significant negative correlations with temperature in 76.3% of the study area (P < 0.05) and significant positive correlations with annual precipitation in 34.1% of the study area (P < 0.05). Herbaceous and evergreen needle-leaf tree cover in the study area were both correlated negatively with temperature (r=-0.889, P < 0.01; r=-0.863, P < 0.01) and were not correlated with annual precipitation.