Cross-sectional studies suggested that consumption of sulfur amino acids (SAAs) including methionine and cysteine is associated with a higher risk of type 2 diabetes (T2D) in humans and T2D-related biomarkers in animals. But, whether higher long-term SAA intake increases the risk of T2D in humans remains unknown. We aimed to investigate the association between long-term dietary SAA intake and risk of T2D. We analyzed data collected from two different cohorts of the Framingham Heart Study, a long-term, prospective, and ongoing study. The Offspring cohort (1991-2014) included participants from fifth through ninth exam, and Third-Generation cohort (2002-2011) included participants from first and second exam. After excluding participants with a clinical history of diabetes, missing dietary data or implausible total energy intake, 3,222 participants in the Offspring and 3,205 participants in the Third-Generation cohort were included. Dietary intake was assessed using a validated food frequency questionnaire. The relations between energy-adjusted total SAA (methionine and cysteine) intake or individual SAA intake (in quintiles) and risk of incident T2D were estimated via Cox proportional hazards models after adjusting for dietary and non-dietary risk factors. Associations across the two cohorts were determined by direct combination and meta-analysis. During the 23 years follow-up, 472 participants reported a new diagnosis of T2D in the two cohorts. In the meta-analysis, the Hazard Ratios (95% Confidence Interval) of T2D comparing the highest with the lowest intake of total SAAs, methionine, and cysteine were 1.8 (1.3, 2.5), 1.7 (1.2, 2.3), and 1.4 (1.0, 2.1), respectively. The association of SAA intake with T2D was attenuated after adjusting animal protein intake in sensitivity analyses. Our findings show that excess intake of SAA is associated with higher risk of T2D. Dietary patterns that are low in SAA diet may help in preventing T2D.