【Objective】Soil heavy metal contamination in mining areas has become an important environmental concern, particularly in arid regions where fragile ecosystems and intensive anthropogenic activities accelerate pollutant accumulation. While environmental magnetism has emerged as a cost-effective proxy for monitoring heavy metals, conventional statistical methods, such as simple correlation analysis, often fail to decipher the complex driving directions and causal mechanisms between magnetic parameters and pollutants. This is especially severe in areas with heterogeneous emission sources. Therefore, this study aimed to characterize the spatial distribution of heavy metals (Zn, Pb, As, Ni, Cr, and Fe) and to elucidate their causal linkages with magnetic mineralogical proxies.【Method】This study focuses on the Qiquanhu coal-mining area in the Turpan Basin, Xinjiang, China. A total of 330 soil samples were collected from 165 sites at two depths (0–10 cm and 10–20 cm) across the study area. Magnetic susceptibility (χ), saturation isothermal remanent magnetization (SIRM), and anhysteretic remanent magnetization (ARM) were measured, alongside the calculation of grain-size- dependent ratios (χARM/χ and χARM/SIRM). The concentrations of six heavy metals were determined using inductively coupled plasma mass spectrometry (ICP–MS). To quantitatively apportion pollution sources, the absolute principal component scores–multiple linear regression (APCS–MLR) receptor model was employed. The geographical convergent cross mapping (GCCM) method was utilized to identify the bidirectional causal relationships and driving strengths between magnetic parameters and heavy metal concentrations based, overcoming the limitations of traditional linear correlation.【Result】The results demonstrated that heavy metal enrichment in the study area was characterized by distinct multi-source superposition. The concentrations of Pb, Zn, and Fe were significantly higher in proximity to industrial facilities, showing a strong coupling with magnetic concentration-dependent parameters (χ, SIRM, and χARM). This suggested that high-temperature smelting activities released coarse-grained magnetic spherules that co-precipitated with these metals. In contrast, As and Ni were primarily associated with fine-grained magnetic particles generated during coal mining and transportation, exhibiting a sensitive response to χARM/χ and χARM/SIRM. However, Cr enrichment showed localized heterogeneity, primarily linked to the accumulation of mining solid waste. The APCS–MLR model identified three primary pollution factors accounting for 86.2% of the total variance, representing industrial smelting, coal-related activities, and lithogenic/waste sources, respectively. GCCM analysis further confirmed that magnetic concentration parameters exerted a strong causal drive on Pb and Zn, while grain-size-sensitive ratios served as robust causal indicators for As and Ni, effectively reducing the risk of misinterpretation inherent in simple correlation.【Conclusion】This study confirms that environmental magnetic parameters can effectively fingerprint the source-specific distribution of heavy metals in arid coal-mining soils. The integration of environmental magnetism, APCS–MLR source apportionment, and GCCM causal analysis provides an advanced diagnostic framework for pollution identification. These findings offer critical scientific evidence for targeted environmental management, risk mitigation, and the development of magnetic-based monitoring protocols in complex polymetallic contaminated regions.