Abstract: The effects of different environmental factors on the spatial variation of soil organic carbon are scale-dependent. Based on soil profile sample data collected from the soil series survey in Hunan Province，soil organic carbon content of the surface layer was obtained by normalizing the soil thickness. The correlations between soil organic carbon content and the influencing factors at different representation scales in were analyzed. The soil organic carbon contents at the sampling scale in Hunan Province was predicted using mathematical statistics and multivariate empirical mode decomposition (MEMD) methods. The purpose of the present study was to investigate the multi-scale spatial relationships between soil organic carbon content and the influencing factors. The results were as follows：(1) MEMD decomposed soil organic carbon into six intrinsic mode functions (IMFs) and one residual. The dominant scale was IMF1，followed by IMF2，and IMF3. The sum of variance contribution rates of these three scales to soil organic carbon accounted for more than 70%. The main representation scales of soil organic carbon and its influencing factors were 133 km (IMF1)，188 km (IMF2) and 355 km (IMF3). (2) The multi-scale analysis indicated that the relationship between soil organic carbon and elevation was mainly manifested at large scales. The correlation between soil organic carbon and its influencing factors became significantly stronger with an increasing scale. (3) The accuracy of soil organic carbon prediction based on the MEMD method was higher than that of using the multiple stepwise regression fitting method，which indicated that multi-scale analysis was more comprehensively in revealing the complex relationship between organic carbon and the influencing factors. The results of this study provided a theoretical basis for the planning and management of soil organic carbon and the improvement of soil quality at different scales.