This paper reports a comprehensive analysis on farm production and resources sustainability in Northwest China during 1978 to 1999. Besides the five provincial regions (Shaanxi, Gansu, Ningxia, Qinghai, and Xinjiang), the upper and middle parts of the Yellow River Basin were also included as a nested area for detailed evaluation. The gross value of farm production was chosen to be the dependent variable in data analysis. Independent variables include farm inputs of irrigation ratio, agricultural labor, fertilizer application, and farm machinery use. In the study period, farm production increased four- to five-fold in the regions, strongly associated with the increased application of technological inputs of fertilizer and machinery. Multivariate regressions were performed based on Cobb-Douglas and Frontier production functions between farm production and inputs. Based on the function parameters, the growth of farm production was decomposed into the contributions by farm inputs and total factor productivity. 10-20% of the growth in farm production was attributed to the increased irrigation ratio in Shaanxi, Gansu, and Ningxia, where the irrigation ratio was relatively lower than other regions and increased significantly during the study period. Productivity in irrigated farmland was found to be over 10 times higher than that of rainfed areas in the regions. For the provincial regions, technological input of fertilizer and machinery contributed to more than half of the growth in farm production in the period of 1978-1998. For the upper and middle parts of the Yellow River Basin, total factor productivity accounted for two-thirds of the farm output growth after 1985. In late 1990s, the regress of total factor productivity caused a significant reduction in farm production due to the drop in grain price and basin-wide drought. Total factor productivity in the basin was decomposed as two components of technological progress and relative technological efficiency. The relative technological efficiency was below 70% with a decreasing trend during 1980 to 1999. To integrate hydrological and economic processes in this study, a spatial model was developed to predict water supply and demand. Some preliminary results of this model were also presented.