The assessment of food provisioning services of ecosystems in Three-gorge areas is helpful for better understanding the function of ecosystems in local human well-beings. In this paper, process-based models are used to assess the potential food provisioning services derived from agriculture ecosystems and grassland ecosystems, a modifying model along with normal woodlands and a set of modifying coefficients is built to assess the potential food from woodland ecosystems. A set of power regression models based on environment factors are built to estimate the potential fish production from water ecosystems. Land cover data stemmed from Landsat TM images, as well as other natural and social-economic data in 1km resolution such as temperature, precipitation, and DEM, are used to support these assessment models. It shows that the four ecosystems in Three-gorge areas can provide 85.98×10 12 calories heat, 2.49 billion kilograms protein and 823.4 million kilograms fat. Human carrying capacity model under the balance nutrition pattern is built in this paper, which results in two key findings: ecosystems in Three-gorge areas can feed 45.92 million people under wealthy living standard which is 1.53 times of the current population, and the sustainable population is from 9.69 to 36.23 million under that living standard. Multi-scale population pressure model is built to calculate the population pressure index in Three-gorge areas. The grain for green pressure index, a multivariate linear weighed model, is used to determine the spatial distribution of farmland fit for grain for green and fit for protecting.