Florfenicol (FF) is used in cattle to treat respiratory diseases but could result in tissue residues. This study aimed to develop a population physiologically based pharmacokinetic (PBPK) model to predict the concentrations of FF and its metabolite, florfenicol amine (FFA), in cattle after four different routes of administration, and to calculate and compare the withdrawal intervals (WDIs) with approved withdrawal times based on different marker residues and their MRLs or tolerances. A flow-limited PBPK model including both FF and FFA sub-models were developed with published data using acslXtreme. This model predicted FF and FFA concentrations in tissues and plasma/serum after intramuscular or subcutaneous administration. Based on the model, the WDIs of 46 and 58 days were calculated to ensure that total residue concentrations (FF + FFA) in 95th percentile of the population after intramuscular and subcutaneous administration were below the MRL, respectively. WDIs were calculated as 44 and 47 days to ensure that FFA concentrations after intramuscular and subcutaneous administration fell below tolerances in 99th percentile of the population, respectively. WDIs were longer than the corresponding label in China, US, and EU. This model provides a useful tool to predict tissue residues of FF and FFA in cattle to improve food safety.