Objective: To explore the association between hematocrit level and risk of incident hypertension. Method: Subjects who participated at least two times routine health check-up in Health Management Center of Shandong Province Hospital between January 2005 and January 2010 were eligible for inclusion. After excluding participants with known hypertension and other related diseases, a prospective cohort with 20 606 subjects (female: 8 218, male: 12 388) was established. Cox-proportional hazard model was used to assess the association between hematocrit and the development of hypertension for female and male respectively. Results: During the 51 352 person-years of follow-up, newly developed hypertension was confirmed in 3 695 cases. For female, the age-adjusted hazard ratios (95%CI) for incident hypertension were 1.00 (reference), 1.06(0.86-1.29), 1.37(1.14-1.65), 1.60(1.34-1.92), respectively (P for trend<0.000 1) through the 3 quartiles of hematocrit levels.After adjusting multiple factors (age, smoking, drinking habit, physical activity, body mass index(BMI), systolic blood pressure, fasting blood glucose, serum creatinine, high-density lipoprotein, gamma-glutamyl transferase, white blood count), the hazard ratios (95%CI) were 1.00(reference), 1.05(0.85-1.29), 1.25(1.03-1.51), and 1.22(1.00-1.48), respectively (P for trend=0.016 9). For male, the hazard ratio (the highest vs. the lowest hematocrit level) after adjusting age or age and life style factors (smoking, drinking habit, physical activity) was 1.23 (1.11-1.37), 1.21(1.09-1.35), respectively. Other analyses of relationship hematocrit with incident hypertension were not statistically significant in male. Conclusions: Higher hematocrit level is associated with higher risk of incident hypertension, especially in female.