Abstract Background We investigated the association between younger-onset type 2 diabetes, duration of diabetes, and cancer risk based on data from the Da Qing Diabetes Prevention Outcome Study (DQDPOS). Methods The analysis recruited 620 younger-onset (age≤50 years) and 649 older-onset (age>50 years) patients with type 2 diabetes, and 310 younger non-diabetes controls (age≤50 years). Multiple regression analysis was used to test the influence of younger-onset diabetes and duration of diabetes on the long-term risk of cancer. Results The annual incidence of all cancer among the non-diabetes, younger-, and older-onset type 2 diabetes was significantly different (3.7, 5.5, and 4.0/1000 person-years, respectively). The standard Cox analysis revealed that the patients with younger-onset diabetes had a significantly higher risk of cancer than those with older-onset diabetes (hazard ratio [HR]:1.81; 95% confidence interval [CI]:1.20–2.73) and younger non-diabetic controls (HR:2.43; 95% CI:1.34–4.41) after adjustment for diabetes duration and other confounders. Stepwise general linear regression model analysis revealed that a longer diabetes-free time was associated with longer lifetime cancer-free years (partial R2 = 0.36, p