Abstract: The study was aimed to establish a liquid chromatography-tandem mass spectrometric method for the determination of the duloxetine concentration in rat plasma, and compare the pharmacokinetics in normal and diabetes mellitus rat models. Diazepam was used as an internal standard. The separation was achieved on a Waters Xterra^® RP18 column (100 mm × 4.6 mm, 3.5 μm) with a mobile phase consisting of methanol -0.3% formic acid containing 5 mmol·L^-1 ammonium acetate (75:25) at the flow rate of 0.6 mL·min^-1. Electrospray ionization source was applied and operated in the positive multiple reaction monitoring mode. A good linearity of duloxetine was obtained in the concentration range of 10-5 000 ng·mL^-1. The rat models of diabetes mellitus were established by intraperitoneal injection of streptozotocin. The same dose of duloxetine (40 mg·kg^-1) was given by intragastric administration to the normal and diabetic rats. Blood samples were collected from the orbital venous plexus to determinate duloxetine concentration in the plasma. The pharmacokinetic parameters were calculated by DAS software. Statistical analysis was performed by SPSS software. The major pharma-cokinetic parameters of diabetes group were as follows: C_max was 1 185 ± 190.0 ng·mL^-1; AUC_0-∞ was 8 398 ± 1 835 ng·mL^-1·h; t_max was 1.6 ± 0.4 h; t_1/2z was 3.6 ± 0.9 h. The major pharmacokinetic parameters of normal group were as follows: C_max was 368.1 ± 40.7 ng·mL^-1; AUC_0-∞ was 4145 ± 640.1 ng·mL^-1·h; t_max was 1.6 ± 0.3 h; t_1/2z was 4.1 ± 0.8 h. The results of pharmacokinetic experiments suggest that the exposure amount of duloxetine in diabetic rats is twice higher than that in normal rats.