To examine the associations of tea consumption (both frequency and type) with (1) prediabetes and diabetes and (2) urinary glucose and sodium excretion in Chinese community-dwelling adults. In 1923 participants (457 with diabetes, 720 with prediabetes, and 746 with normoglycaemia), the frequency (occasional, frequent, daily, or nil) and type (green, black, dark, or other) of tea consumption were assessed using a standardized questionnaire. Morning spot urinary glucose and urine glucose-to-creatinine ratios (UGCRs) were assessed as markers of urinary glucose excretion. Tanaka's equation was used to estimate 24-h urinary sodium excretion. Logistic and multivariate linear regression analyses were performed. Compared with non-tea drinkers, the corresponding multivariable-adjusted odds ratios (ORs) for prediabetes and diabetes were 0.63 (95% confidence interval [CI] 0.48, 0.83) and 0.58 (95% CI 0.41, 0.82) in participants drinking tea daily. However, only drinking dark tea was associated with reduced ORs for prediabetes (0.49, 95% CI 0.36, 0.66) and diabetes (0.41, 95% CI 0.28, 0.62). Dark tea consumption was associated with increased morning spot urinary glucose (0.22 mmol/L, 95% CI 0.11, 0.34 mmol/L), UGCR (0.15 mmol/mmol, 95% CI 0.05, 0.25 mmol/L) and estimated 24-h urinary sodium (7.78 mEq/day, 95% CI 2.27, 13.28 mEq/day). Regular tea consumption, especially dark tea, is associated with a reduced risk of dysglycaemia and increased urinary glucose and sodium excretion in Chinese community-dwelling adults.