In this paper, an empirical Bayes methodology was used to determine the pharmacokinetic profile of sodium tungstate in beagle dogs after multiple oral dosing using the P-PHARM computer program. The population estimation algorithm used in P-PHARM is an EM-type procedure. Sodium tungstate was administered orally, three times a day, (i) for 11 days (21 and 42 mg/kg per day) to 18 dogs (nine males and nine females) and (ii) for 13 weeks (15, 30 and 60 mg/kg per day) to 28 dogs (14 males, 14 females). Six other dogs received the compound intravenously (25 and 50 mg/kg). Plasma concentration profiles versus time were compatible with a two-compartment model and first-order kinetics. After oral administration, F (0.61±0.086 vs. 0.48±0.093), and normalized (to a 7-mg/kg dose of sodium tungstate) AUC (54±8.4 vs. 41.2±8.5 mg/l×h), C max (10.6±0.49 vs. 8.5±0.57 μg/ml) and C min (3.04±0.23 vs. 2.04±0.22 μg/ml), were higher in male than in female dogs. However, the introduction of the gender in the final model did not contribute statistically to an improvement of the fit of the population pharmacokinetic model. In males, t 1/2 elimination averaged 3.1±0.56 vs. 2.6±0.18 h in females. The duration of treatment did not modify statistically the pharmacokinetic parameters. After repeated multiple oral administration of 15–60 mg/kg per day of sodium tungstate, tungsten plasma concentrations increased in proportion to dose. No dose-dependent changes in pharmacokinetic parameters occurred.