Global-level models have consistently found that climate change will increase the risk of hunger. The climate-undernutrition relation is complex, and choices must be made about what is brought into focus, with these choices drawing attention to particular causes and solutions. A critical overview of the literature showed, however, that all previous models had adopted one general conceptualisation: less or lower quality food means more hunger. This leaves much unexplored. The central idea of this thesis is that when faced with complex health problems, a fuller understanding may be gained by developing multiple models, each adopting a different perspective. The thesis aims to develop a series of global-level climate-undernutrition models, with the insights from one model guiding the development of the next. These models are presented as four research papers. Papers 1 and 2 adopt a ‘crop productivity’ perspective to quantify stunting. The results suggest that future socioeconomic conditions will play a larger in role in shaping stunting than climate change. Thus, Paper 3 places food production in the background and asks how climate change may impact on stunting via its impacts on two socioeconomic factors: incomes and food price. The results imply that slowly rising food prices lead to decent farm incomes, which may reduce the risk climate change poses to nutrition in rural areas. Producer-consumer farmers, however, were not directly represented. Given this, Paper 4 assesses the health-related implications for rural populations of producer-consumer households practising different styles of farming in the global food system under climate change. The results suggest that how farming is done – whether more entrepreneurial- or peasant-like – will impact on future nutrition and the conditions that support rural health. Collectively, the research papers demonstrate the utility of a multiple model-approach to complexity, and the benefits of drawing on a range of theories when building models.