OBJECTIVE： To study the dermal pharmacokinetic difference of triptolide in normal and diabetic rats， and to provide reference for rational drug use in the clinic. METHODS： 12 Wistar rats were randomly divided into normal group and diabetic model group （0.1％ streptozotocin intraperitoneally）， with 6 rats in each group. Both group were given Triptolide cream 0.5 g to abdominal skin， and dialysate was collected by microdialysis every 30 min for consecutive 12 h. Subcutaneous concentration was detected by HPLC-MS， and subcutaneous concentration-time curves were analyzed and compared between two groups， and Winnonlin 5.0.1 software was used to calculate pharmacokinetic parameters. RESULTS： The pharmacokinetic parameters of normal group and diabetic model group were that cmax  were （1.54±0.37） and （5.12±1.34） μg/ml； tmax were （7.32±0.24） and （6.21±0.35） h； AUC0-12 h were （12.65±4.64） and （37.43±5.23） μg·h/ml， with statistical significance （P＜0.05）. CONCLUSIONS： The change of dermal structure caused by diabetes can increase percutaneous penetration amount of triptolide in rats， and drug dosage should be reduced according to circumstances so as to reduce side effects.