Understanding the spatial distribution of soil properties in Arctic landscapes can improve our knowledge of carbon storage and the impacts of climate change. This study utilized a range of covariates, including organisms, climate, topography, soil, geology, and land use types, to map and identify key variables responsible for the spatial distribution of soil organic carbon (SOC) and pH values over a 14.000 km2 area in northern Russia. To this end, we employed three machine learning methods: Random Forest (RF), K-Nearest Neighbor (KNN), and Cubist. Our results indicated that the RF outperformed the KNN and Cubist algorithms for both soil parameter predictions and that organisms and climate data (surface temperature and cloud cover) were identified as key variables, with the highest contribution to the models. The generated maps showed that the highest SOC concentrations were associated with alluvial and peat soils, whereas the lowest content was predicted in the mountains. The more acidic soils were concentrated in the flat part of the region, whereas the more alkaline soils were located in the foothills and mountains. This research is important in the context of climate change, as the northern regions are critical for the global carbon cycle and play an essential role in regulating the Earth's climate. The information obtained from this study can aid in predicting future carbon fluxes, mitigating the impact of climate change, and promoting sustainable land management practices.