【Objective】The thinning of the black soil layer is one of the primary issues in black soil degradation, and accurate identification of its thickness and spatial variations holds significant importance for the protection and sustainable utilization of black soils. This study aimed to integrate ground penetrating radar (GPR) and variational mode decomposition (VMD) technologies to develop methods for in-situ identification and spatial prediction of black soil layer thickness.【Method】Three typical slope surfaces with thick, medium, and thin black soil layers in Nenjiang, Heilongjiang Province, were selected as the study sites. The 700 MHz GPR was used to detect the black soil layer thickness along the slope surfaces and validated the GPR identification accuracy using data from soil drilling and profile surveys. The VMD technology was applied to decompose the GPR signals, selected key modal components through spectrum analysis, and combined this with envelope analysis to precisely identify the black soil layer thickness. Furthermore, the sparse sampling (lateral ratio of 50:1 and longitudinal ratio of 100:1) and natural neighborhood interpolation was applied to predict the spatial distribution of black soil layer thickness at the slope scale.【Result】The GPR-VMD method effectively separated noise from the signals, resulting in root mean square errors for black soil layer thickness detection at the profile scale ranging from 2.03 to 6.86 cm. The average root mean square errors for the thick, medium, and thin black soil layers were 6.26, 4.30, and 3.73 cm, respectively. Overall, across all validation points, the overall detection root mean square error was approximately 6.08 cm, the overall prediction error was controlled within 9.7%, and the overall coefficient of determination (R2) reached 0.94. At the slope scale, the spatial predictions of black soil layer thickness showed significant correlations with measured data (R2 > 0.87), enabling efficient, non-destructive, and continuous detection of black soil layer thickness. The generated spatial distribution maps revealed associations between black soil layer thickness and factors such as slope position and terrain, illustrating gradient changes and spatial differentiation patterns induced by erosion processes.【Conclusion】The integration of GPR and VMD methods achieves efficient, non-destructive, and high-precision in-situ identification and spatial prediction of black soil layer thickness in typical black soil regions of Northeast China. This study provides important methodological and technical support for revealing the spatial-temporal variations and underlying mechanisms influencing black soil layer thickness.