BACKGROUND: Ultra-processed foods account for more than 50% of daily calories consumed in several high-income countries, with sales of ultra-processed foods soaring globally, especially in middle-income countries. The objective of this study is to investigate the association between ultra-processed food (UPF) consumption and risk of type 2 diabetes (T2D) in a UK-based prospective cohort study. METHODS: Participants of the UK Biobank (2007-2019) aged 40-69 years without diabetes at recruitment who provided 24-h dietary recall and follow-up data were included. UPFs were defined using the NOVA food classification. Multivariable Cox proportional hazards regression models were used to evaluate the association between UPF consumption and the risk of T2D adjusting for socio-demographic, anthropometric and lifestyle characteristics. RESULTS: A total of 21,730 participants with a mean age of 55.8 years and mean UPF intake of 22.1% at baseline were included. During a mean follow-up of 5.4 years (116,956 person-years), 305 incident T2D cases were identified. In the fully adjusted model, compared with the group in the lowest quartile of UPF intake, the hazard ratio for T2D was 1.44, 1.04-2.02 in the group with the highest quartile of UPF consumption. A gradient of elevated risk of T2D associated with increasing quartiles of UPF intake was consistently observed (p value for trend < 0.028). A significantly increased risk of T2D was observed per 10 percentage points increment in UPF consumption ([adjusted HR]: 1.12, 95% confidence interval [CI]: 1.04-1.20). CONCLUSIONS: Our findings demonstrate that a diet high in UPFs is associated with a clinically important increased risk of T2D. Identifying and implementing effective public health actions to reduce UPF consumption in the UK and globally are urgently required.