Expansion of groundwater-fed irrigation has enhanced agricultural productivity, while simultaneously causing large reductions in groundwater storage. Hydro-economic modelling provides a valuable tool for evaluating trade-offs between agriculture and aquifer sustainability. However, a central argument in this thesis is that the reliability of integrated analysis of groundwater systems may be affected by the failure to consider adequately the variables influencing field-level irrigation decisions and the dynamic aquifer responses to those choices. This thesis addresses this limitation of existing research through three main contributions to the literature. In Chapter 2, a new model of farmers' individual field-level groundwater-fed irrigation decision making is developed. Innovations include the explicit consideration of the role of soil moisture in intraseasonal irrigation scheduling, and evaluation of the impacts of weather variability and well yield on crop production. Application to a case study in the High Plains region of the United States demonstrates that low well yields limit significantly irrigated area and demand for groundwater. Furthermore, it is shown that this important behavioural response can not be captured by existing models that do not consider instantaneous supply constraints imposed by well yield. In Chapter 3, a combination of empirical analysis of observation data, and numerical simulations using the model developed in Chapter 2, are conducted to generate insights about the feedbacks between groundwater and agriculture. Reductions in well yield are shown to have large negative impacts on the productivity of irrigation whereas, contrary to previous assumptions, the response of irrigation behaviour to increased groundwater pumping lifts is found to be minimal. Importantly, this result highlights the need for integrated models to consider a broader range of factors influencing the sustainability of agricultural groundwater use. In Chapter 4, a novel hydro-economic modelling framework is presented that integrates the model of farmers' individual irrigation decision making with a distributed multi-scale groundwater flow model. The model extends the scope of existing research by using pump head-capacity curves to quantify the effect of changing aquifer storage on well yields. Advantages of the modelling framework for policy analysis in coupled agricultural groundwater systems are demonstrated. Specifically, it is shown that the long-term sustainability of irrigated agriculture is influenced by intraseasonal and interannual changes in well yield that depend on famers' pump choice and expectations about future dynamic aquifer responses. Moreover, the model demonstrates that managing the rate of well yield decline may have positive long-term impacts on both agricultural production and farmer welfare that have not been recognised in previous research.