This thesis investigates the role of food preferences and dietary patterns in shaping multimorbidity-related health outcomes, using data from the UK Biobank. This thesis investigates the role of food preferences and dietary patterns in shaping multimorbidity-related health outcomes, using data from the UK Biobank. A literature review conducted as part of this research confirmed that healthy dietary patterns are consistently associated with lower risks of cardiovascular disease, type 2 diabetes mellitus, and colorectal cancer—highlighting the importance of dietary behaviour in chronic disease prevention. Despite this, the review also revealed that food preference questionnaires remain largely unused in studying diet–disease associations, even though they offer valuable insights into individual-level dietary inclinations. To address this gap, the thesis applied unsupervised machine learning methods, three distinct food preference profiles were identified and characterised as putative Health-conscious, Omnivore, and putative Sweet-tooth. These profiles were found to be significantly associated with specific disease risks and exhibited unique blood-borne biomarker signatures, suggesting their potential utility in predicting health trajectories. To understand the underlying drivers of food preferences, this study examined a combination of individual and environmental factors. Early-life exposure to sugar-rich diets was shown to influence later-life food preferences and associated health risks. Genetic variation was also found to play a partial role in shaping individual food likes. However, when translating food preferences into actual dietary choices, environmental and social influences appeared to exert a greater impact than taste preference alone. Building on these insights, the thesis proposes a conceptual pipeline for a Digital Health Intervention (DHI), aimed at preventing the incidence of cardiovascular disease. This framework integrates food preference profiling with behaviour change strategies, laying the groundwork for a personalised, data-driven approach to nutrition and health promotion. A key contribution of this work is the identification of three meaningful food preference profiles, each linked to distinct biological and clinical outcomes. These findings provide a strong foundation for developing targeted, personalised nutrition strategies that can contribute to disease prevention and the promotion of optimal health.