To establish an accurate method for the geographical discrimination of Zanthoxylum, this study analyzed differences in characteristic quality indicators including total sanshoamides, volatile oils and sanshools among 106 batches of Zanthoxylum bungeanum and Z. schinifolium from representative production regions in Gansu, Sichuan, Yunnan, Shaanxi and Chongqing provinces. The differences in aroma components were determined by solid-phase microextraction (SPME) using a DVB/C-WR/PDMS fiber, separation on a DB-5MS capillary column (30 m × 0.32 mm, 0.25 μm) and gas chromatography (GC) using an electron ionization (EI) source. Through flavoromics approaches and multivariate data analysis, six characteristic differential aroma components (sabinene, myrcene, D-limonene, eucalyptol, α-pinene, and L-α-pinene) across geographical origins and species were identified. A four-component model for the geographical discrimination of Zanthoxylum was developed based on these six aroma components and validated using 28 blind samples, obtaining 100% accuracy. The model has been successfully applied in the geographical discrimination of Zanthoxylum. This method, overcoming the market over-reliance on traditional sensory evaluation, enables rapid geographical discrimination of Zanthoxylum. This study provides a reference for establishing a Zanthoxylum identification and traceability system.