Background: Pleuropterus multiflorum Thunb. cv. “Heshouwu”(HSW) has been used as a classical material for both medicine and food in China for millennia. Recently, the cultivation region of HSW has shifted from Guangdong to Sichuan, Guizhou, and other regions. The investigation of geographic variation in bioactive metabolite contents and their relationship with soil mineral elements holds academic significance. Objective: This study aimed to investigate the variations in the distribution of active components in HSW across diverse planting regions and their relationship with soil mineral elements. Methods: A reliable quantitative analysis based on ultrahigh-performance liquid chromatography with triple-quadrupole mass spectrometry (UPLC-QQQ-MS) was developed to assess the levels of 15 bioactive metabolites in 60 HSW samples collected from 4 distinct regions. A total of 43 soil mineral elements in corresponding 60 soil samples were quantified by inductively coupled plasma mass spectrometry (ICP-MS). Orthogonal partial least squares-discriminant analysis (OPLS-DA), heatmap analysis, Pearson correlation analysis, and random forest (RF) regression were conducted based on the above quantitative data. Results: The content of stilbene glycosides displayed a wider range of variation compared with emodin and physcion among different regions. Eight compounds were screened as the differential metabolites in HSW samples from various sources using the supervised OPLS-DA analysis. Among these, 2 important functional compounds including physcion and 2, 3, 5, 4′-tetrahydroxystilbene-2-O-(6″-O-ace-tyl)-glucoside (THSG-5) are the most abundant in HSW samples from Deqing, a geographical indicative production region. Pearson correlation analysis indicated that the impact of soil mineral elements on the levels of stilbene glycosides is greater compared to that on anthraquinones. A negative correlation was observed between the levels of elements Na, Zn, Ba, Ti, and 2, 3, 5, 4′-tetrahydroxysilbene 2-O-glucoside (THSG-1). Conversely, a positive correlation was found between the contents of elements Na, Ce, Ti, and physcion and THSG-5, 2 components that exhibited higher levels in Deqing. Furthermore, an RF algorithm was employed to establish an interrelationship model, effectively forecasting the abundance of the majority of differential metabolites in HSW samples based on the content data of soil mineral elements. Conclusions: The variation of stilbene glycosides is wider than emodin and physcion in HSW. The levels of metabolites in HSW samples are influenced by soil mineral elements, with stilbene glycosides being more susceptible to such influences compared to anthraquinones. Specifically, THSG-1 shows a negative association withmost soil mineral elements, notably Na, Zn, Ba, and Ti, whereas the content of physcion displays a positive correlation.