Purpose: Tumor hypoxia is correlated with treatment failure in radiotherapy. The hypoxic fraction of a tumor can be difficult to define using non‐invasive imaging methods such as Positron Emission Tomography (PET). The purpose of this study is to use tracer kinetic modeling techniques to increase the accuracy and precision of hypoxic volume (HV) quantification with 18F‐fluoromisonidazole (18F‐FMISO) PET imaging. Methods: Ten male BALB/c mice with EMT‐6 tumors were selected for imaging. Anaesthetized animals were injected with 18F‐FMISO and imaged using whole‐body 120‐min dynamic PET and Computed Tomography (CT). Data from dynamic 18F‐FMISO scans for 3 mice were fit to a 2‐compartment irreversible 3‐rate constant model (K1, k2, k3) and a PATLAK model (Ki). Different thresholds were applied to the resultant Ki and k3 rate constants generated on a voxel‐by‐voxel basis to calculate the HV for each tumor. These kinetic‐derived HVs were compared to HVs generated using conventional tumor‐to‐blood (TBR) ratios. Results: Kinetically‐derived HV (using the k3 tracer binding constant and the rate of tracer influx, Ki), calculated on a voxel‐by‐voxel basis, showed an increase in HV of up to 38% over the TBR method depending on the assumed threshold. Moreover, the PATLAK kinetic model using a Ki threshold may provide better delineation and calculation of the tumor HV than a 2‐compartment pharmacokinetic model. Conclusion: For PET imaging with 18F‐FMISO, pharmacokinetic modeling can improve the accuracy and precision of tumor hypoxia quantification over that obtained with tumor‐to‐blood ratios. In particular, the PATLAK method to define a tumor HV may be superior due to a reduction in noise and to better HV localization. Ongoing studies will correlate tumor HVs with Eppendorf electrode measurements and carbonic anhydrase IX staining of histologic sections. These studies could lead to improved techniques for identifying patients likely to benefit from therapies designed to overcome hypoxic radioresistance.