A revised version of an atmosphere-vegetation interaction model AVIM is presented and applied to the Northeast China Transect (NECT), an ecological transect over the arid/semi-arid area of China, to study the interaction between land surface physical processes and the physiological processes. The model reasonably simulates the physical fluxes and net primary productivity (NPP) along NECT. A 41-year simulation along NECT reveals the following. (1) Moisture is a major factor controlling the spatial distributions and temporal variation of NPP and leaf area index (LAI) along NECT. It is found that the effect of temperature-induced (or radiation-induced) variation in the moisture condition of the vegetation exceeds the effect of temperature (or radiation) itself on NPP and LAI. (2) The spatial distribution of vegetation and related parameters, such as albedo and canopy resistance, is crucial to the simulated distributions or the gradients of the physical fluxes. The simulated interannual variability of surface physical fluxes is strongly related to the interannual variation of the vegetation. (3) Related to global warming, there are long-term trends in the NPP and LAI of natural vegetation along NECT. At the early part of the growing season (April), the trends of NPP and LAI are positive mainly owing to the lengthening of growing days in April; however, the trends are negative during the summer (the main growing season) due to the unfavorable effects on the moisture condition of the vegetation by the warming.