Forest ecosystems attract a continuous attention because of their importance for mitigating environmental change. Policy making in the fields of sustainable forest management and agendas related to climate change drives the research on forest carbon budget and its interpretation. The recent adoption of the Kyoto Protocol further spurs the need of a sound understanding of carbon-related processes. There is a big challenge to accurately estimate the contribution of forest ecosystems to the global carbon cycle,as it is difficult to directly measure carbon pools or fluxes over large area. Some models are available to address the issues of carbon budget,and they may be classified into several ways. In terms of their capability to include the causality,the models can be either empirical regression models,or so-called process models including eco-physiological processes describing ecosystem functioning in terms of key processes with their interactions. The adopted simulation system may include all kinds of combinations of approaches to fulfill the modeling goals and create the specific model hybrids. In any case,the models developed based on the generalized approach to simulate ecosystem development must be parameterized for the target ecosystems. Regarding the estimation of carbon dynamics,the use of process-based ecosystem models is of particular interest because this approach allows not only to estimate the carbon budget under various environmental conditions, but also to interpret and quantify the possible causes of carbon stock change along with the environmental change. The application of ecosystem models,however,encounters some specific difficulties in dryland environment,which is characterized by its climatic and human-induced features. After identifying the proper ecophysiological parameters used in Biome-BGC model,in this study the NPP (net primary productivity) and NEP (net ecological productivity) in the Tianshan Mountains were estimated. Furthermore,the responses of NPP /NEP to interannual climate change during the period of 1959-2009 and climate change scenarios in the future were modeled. Results showed that the average annual total NPP and NEP were 547. 97 g·m~(-2)·a~(-1) and 61. 24 g·m~(-2)·a~(-1) respectively, which revealed that the forests in the Tianshan Mountains perform as the carbon sinks. NPP and NEP were sensitive to the change of precipitation and the increase of air temperature. Under a scenario of 4 ℃ temperature increase, NPP was slightly improved,but NEP was decreased as soil respiration was boosted up. Drought stress was not significant in the study area as derived from the results.