Sex-specific differences in pharmacokinetics are increasingly recognized as clinically relevant, yet most population pharmacokinetic models for antibiotics have been developed in predominantly male populations with limited transferability to female patients. This study aimed to develop and externally evaluate a population pharmacokinetic model of piperacillin specifically for critically ill women to support model-informed precision dosing (MIPD). A total of 923 samples from 165 critically ill women (from four previously published studies) receiving piperacillin/tazobactam were analyzed. The data set was randomly split into a training set (80%) and an independent external evaluation set (20%). Non-linear mixed-effects modeling was employed based on the training data using NONMEM, and covariates were identified using stepwise covariate modeling with cross-validation. Prediction accuracy and bias were quantified using the median prediction error (MDPE) and the median absolute prediction error (MDAPE) in an MIPD scenario. Model predictions for future concentrations were generated both a priori and through Bayesian forecasting based on one prior sample. A two-compartment model provided the best fit for the data. Clearance was significantly affected by renal function estimated by the Cockcroft-Gault equation, along with procalcitonin and serum albumin. The external evaluation demonstrated strong predictive performance with very low systematic bias. In the MIPD scenario, a priori predictions exhibited good predictive performance (MDPE < |5.2|%, MDAPE < 33.4%), which substantially improved with Bayesian forecasting (MDPE < |3.2|%, MDAPE < 19.1%). A female-specific population pharmacokinetic model of piperacillin was developed and externally validated, demonstrating robust predictive performance and clear benefits for Bayesian forecasting.