The relationship between juice consumption and type 2 diabetes (T2D) has not been widely evidenced. Our aims were to prospectively evaluate the associations with T2D incidence of: 1) isovolumetric substitution of a water serving/day for one of fruit juice (different types), and of fresh fruit juice for its bottled version; 2) consumption of total, fresh or bottled juice; 3) energy intake from juices. We followed 17,518 adults without T2D at baseline. Beverage consumption was assessed at baseline through a validated food-frequency questionnaire. The outcome was T2D incidence, according to American Diabetes Association's criteria. During a median follow-up of 10.2 years, 142 incident cases of T2D were identified. In substitution models, the substitution of water for bottled juice was associated with a lower T2D incidence, and also if the replacement was done by fresh juice, or especially fresh orange juice HR 0.75 (95% CI 0.57–0.99), 0.65 (95% CI 0.43–0.98) and 0.56 (95% CI 0.34–0.92); respectively. Each additional serving/day of bottled juice was directly associated with T2D incidence HR 1.33 (95% CI 1.01–1.75). No significant association was observed for energy coming for bottled juice HR 1.74 (95% CI 0.94–3.20). Our results suggest that isovolumetric substitution of water or fresh juice for bottled juice was inversely associated with T2D incidence in a long-term prospective study. Thus, these substitutions could be useful to tackle the diabetes epidemic. •A water serving/day instead of bottled juice could decrease diabetes incidence.•The same effect was observed when it is replaced by fresh juice.•Each additional serving per day of bottled juice would increase diabetes risk.•Energy coming from juices would not be associated with diabetes risk.