<p>Reliable authentication of high-value edible oils remains challenging when adulterants are unknown and sample heterogeneity affects model robustness. In this study, gas chromatography-ion mobility spectrometry (GC-IMS) fingerprinting was integrated with an origin-assisted one-class chemometric strategy for non-targeted authentication of camellia oil (CAO) adulteration. Volatile organic compound fingerprints of authentic CAO from seven geographical origins and their adulterated samples were systematically characterized. A two-step framework was developed in which geographical origin was first identified using PLS-DA, followed by origin-specific adulteration authentication using one-class classification algorithms (DD-SIMCA and OC-PLS). This strategy reduced origin-induced variability and improves model reliability compared with conventional one-step modeling. DD-SIMCA achieved authentication accuracies above 94% across all adulteration systems and showed strong sensitivity to low-level adulteration. In addition, PLS regression enabled accurate quantification of adulteration levels for four edible oils (R_p² &gt; 0.96, RPD &gt; 4). These results demonstrate that GC-IMS fingerprinting combined with origin-assisted one-class chemometric modeling provides a rapid, reliable, and non-targeted analytical framework for edible oil authentication, with good potential for extension to other high-value oils.</p>