Objcetive To investigate the relationship between blood β-hydroxybutyric acid and urine ketone in the di-agnosis of diabetic ketosis (DK). Methods Peripheral blood β-hydroxybutyric acid and urine ketone were detected when the peripheral blood glucose was more than 13.9 mmol/L in patients with diabetes. Results (1) In 81 diabetes pa-tients with blood glucose more than 13.9 mmol/L, the incidence of DK was 13.58% and the incidence of diabetic ke-toacidosis (DKA) was 9.88%. (2) The peripheral blood glucose was positively correlated with β-hydroxybutyric acid (r=0.330, P=0.003), but it was not correlated with urine ketone. (3) The peripheral blood β-hydroxybutyric acid was posi-tively correlated with urine ketone (r=0.516, P=0.000). (4) In patients with DK or DKA, 5.26%(1/19) of those were with urine ketone(-) or (+-), whereas 36.84% (7/19) of those were with blood β-hydroxybutyric acid less than 1 mmol/L. (5) When urine ketone was used as the reference test for diagnosis of DK, the optimal value of blood β-hydroxybutyric acid was 0.35 mmol/L. Conclusion For missed diagnosis of DK may be happend if blood β-hydroxybutyric acid or urine ketone is used alone, the co-monitoring of blood β-hydroxybutyric acid and urine ketone can reduce the inci-dence of missed diagnosis of DK. The urine ketosis may have existed when the blood β-hydroxybutyric acid is slightly elevated (≥0.35 mmol/L). In the situation, the urine ketone should be tested in order to avoid missed diagnosis of DK.