Efficient use of water and nutrients in crop production are critical for sustainable water and crop production systems. Understanding the role of humans in ensuring water and nutrient use efficiency is therefore an important ingredient of sustainable development. Crop production functions are often defined either as functions of water and nutrient deficiency or are based on economic production theory that conceptualizes production as a result of economic activities that take in inputs such as water, capital and labor and produce crop biomass as output. This paper fills a gap by consistently treating water and nutrient use and human agency in crop production, thus providing a better understanding of the role humans play in crop production. Uptake of water and nutrients are two dominant biophysical processes of crop growth while human agency, including irrigation machine power, land-preparing machine power and human labor force, determine limits of water and nutrient resources that are accessible to crops. Two crops, i.e., winter wheat and rice, which account for the majority of food crop production are considered in a rapidly developing region of the world, Jiangsu Province, China, that is witnessing the phenomenon of rural to urban migration. Its production is modeled in two steps. First water and nutrient efficiencies, defined as the ratios of observed uptake to quantities applied, are modeled as functions of labor and machine power (representing human agency). In the second step, crop yields are modeled as functions of water and nutrient efficiencies multiplied by amounts of water and fertilizers applied. As a result, crop production is predicted by first simulating water and nutrient uptake efficiencies and then determining yield as a function of water and nutrients that are actually taken up by crops. Results show that modeled relationship between water use efficiency and human agency explains 68% of observed variance for wheat and 49% for rice. The modeled relationship between nutrient use efficiency and human agency explains 49% of the variance for wheat and 56% for rice. The modeled relationships between yields and actual uptakes in the second step explain even higher percentages of observed the variance: 73% for wheat and 84% for rice. Leave-one-out cross validation of yield predictions shows that relative errors are on average within 5% of the observed yields, reinforcing the robustness of the estimated relationship and of conceptualizing crop production as a composite function of bio-physical mechanism and human agency. Interpretations based on the model reveal that after 2005, mechanization gradually led to less labor being used relative to machinery to achieve same levels of water use efficiency. Labor and irrigation equipment, on the other hand, were found to be complimentary inputs to water use efficiency. While the results suggest interventions targeting machinery are most instrumental in increasing wheat productivity, they may exasperate rural – urban migration. Policy strategies for alleviating rural-urban migration while ensuring regional food security can nonetheless be devised where appropriate data are available.