<p>Rapid detection of leaks in water distribution systems is conducive to saving resources and protecting the environment. Ground-penetrating radar (GPR) is an efficient and economical nondestructive testing technique that can realize the rapid detection of pipeline leaks. However, because the diffraction characteristics of the leakage area are not obvious, it is difficult for the traditional GPR to accurately extract the scattering information of the leakage area in the complex buried environment. This paper proposes a fast pipeline leakage detection method based on eigenvalue decomposition with full polarimetric ground-penetrating radar (FP-GPR). Aiming at the limitations of the traditional <i>H-α</i> decomposition regarding parameter instability and computational efficiency, we propose more stable and computationally efficient substitution parameters <i>h</i> and <i>q</i>. Based on the roughness state of the interface in the leakage area and the saturation distribution of the soil, we analyze the influencing factors of the polarized rotating field. Numerical simulations validate the workflow of the method, and modeling experiments verify the method's applicability in engineering. The results show that pipe leakage leads to a decrease in <i>α</i> and an increase in <i>H</i>. The <i>H-q</i> and <i>h-q</i> planar classifiers are more favorable for pipeline leakage identification. The method can be adapted to rapidly detect pipe leakage in complex scenarios, which is significant for managing and maintaining urban water distribution systems.</p>