We investigated the association between creatinine to body weight (Cre/BW) ratio and incident diabetes. In this cohort study, 9659 men and 7417 women were follow up mean (SD) 5.6 (3.5) years and 5.4 (3.4) years, respectively. For men, tertile 1 (T1; n = 3176), Cre/BW < 0.01275; tertile 2 (T2; n = 3258), 0.01275 ≤ Cre/BW < 0.0148; and tertile 3 (T3; n = 3225), Cre/BW ≥ 0.0148; and for women, T1 (n = 2437), Cre/BMI < 0.0118; T2 (n = 2516), 0.0118 ≤ Cre/BMI < 0.014; and T3 (n = 2477), Cre/BMI ≥ 0.014. Among them, 362 men and 102 women developed diabetes. The hazard ratios (HRs) of incident diabetes in the T2 group was 0.56 (95% CI 0.44–0.71, p < 0.001) in men and 0.61 (0.38–0.99, p = 0.045) in women and in the T3 group was 0.42 (0.32–0.54, p < 0.001) in men and 0.55 (0.34–0.89, p = 0.014) in women after adjusting for covariates, compared with the T1 group. Moreover, Δ0.001 incremental of Cre/BW is negatively associated with incident diabetes (adjusted HR 0.84, 95% CI 0.80–0.88, p < 0.001 for men and 0.88, 0.81–0.96, p = 0.003 for women). In conclusion, Cre/BW ratio is inversely related to incident diabetes. Checking Cre/BW ratios may predict future diabetes risks.