Reliable origin certification methods are essential for the protection of high-value genuine medicinal material with designated origins and geographical indication (GI) products. Aconiti Lateralis Radix Praeparata (Fuzi), one well-known traditional Chinese medicine and geographical indication products have remarkable efficacy and wide clinical application, with high demand in domestic and international markets. The efficacy and price of Fuzi from different origins vary, and it is difficult for the general public to accurately identify them through traditional experience. The mass spectrometry detection technology based on the plant metabolomics is tedious and lengthy in test sample preparation, complicated in operation, long in detection time, and low in reproducibility. As a sophisticated, green, fast, and low-loss detection technique, infrared spectroscopy is integrated by machine learning to bring new ways for quality regulation and control of traditional Chinese medicines. An analytical method based on mid-infrared spectroscopy combined with a random forest algorithm was developed to verify the geographical origin of authentic herbs and/or GI products. The method successfully predicted and classified three varieties of Chinese GI Fuzi and four varieties of non-GI Fuzi. In this study, an environment-friendly traceability strategy with fast analysis, low sample loss and high precision was used to provide a new strategy for identifying the origin of Fuzi.