A physiologically based pharmacokinetics model was developed to predict tulathromycin concentrations in edible swine tissues. Physiological parameters included volumes of and plasma flows through different tissues which were obtained from the literatures. The tissue/plasma partition coefficient was calculated according to the area method, and the model was validated through a comparison of predicted and observed concentrations. Withdrawal times in different tissues were predicted. The physiologically based pharmacokinetics model presented here provided accurate predictions of the observed concentrations in all tissues. The results showed that the injection site had the longest withdrawal time (21 days), followed by skin together with fat (19 days) and then kidney (10 days), lung (6 days), liver (4 days) and muscle (1 day). A withdrawal time of 21 days was finally predicted for tulathromycin in swine after a single intramuscular injection at 2.5 mg/kg body weight.