This study aimed to establish a population pharmacokinetic (PopPK) model using data from 2 clinical trials of zimberelimab, evaluate the pharmacokinetics (PKs) of zimberelimab, explore the feasibility of 360 mg once every 3 weeks (Q3W) and 480 mg once every 4 weeks (Q4W) as alternative dosage regimens, and analyze the exposure-response relationship of the efficacy and safety of zimberelimab for advanced tumors. The PKs of zimberelimab were described using the 2-compartment model with time-dependent nonlinear elimination. The prediction-corrected visual predictive check was used to evaluate the model's predictive value on blood drug concentrations. In total, 2165 PK observations from 321 participants were included. The PopPK model demonstrated a high level of concordance between the observed data and the predicted values, indicative of a robust fit to the PK data of zimberelimab. The PK variables were similar for the 240 mg once every 2 weeks, 360 mg Q3W, and 480 mg Q4W regimens. No covariates significantly affecting the PK variables in the final model were found. The exposure variables of zimberelimab have no obvious correlations with efficacy and safety, and 360 mg Q3W and 480 mg Q4W are worthy of further study. This study establishes a PopPK model and analyzes the exposure-response relationship of zimberelimab, which helps to explore the potential for alternative dosing regimens and offers a foundation for optimizing therapeutic strategies for advanced cancer patients through simulation-based methods.