Soil moisture serves as a crucial water source for vegetation growth, while vegetation dynamics in turn regulate soil moisture through processes such as evapotranspiration and canopy interception. Quantitatively characterizing this mutual feedback is essential for optimizing hydrological processes and promoting coordinated vegetation-water management. However, the bidirectional interactions between soil moisture and vegetation productivity across different climate zones, vegetation types, and soil depths remain poorly understood. Based on Global Solar Induced Chlorophyll Fluorescence, soil moisture, the average air temperature at 2 m, and solar radiation datasets, this study employed trend analysis, causal inference, and multiple linear regression to explore the Granger causality between vegetation productivity and multi-layer soil moisture under different environmental conditions. Results revealed widespread bidirectional dependence between vegetation productivity and soil moisture across all soil layers, with interaction strength declining from 79.84% in the surface layer (0-10 cm) to 72.04% in the deep layer (100-200 cm). The magnitude of this bidirectional coupling varied significantly across climate zones: dependence peaked in surface soils within temperate regions (82.52%), while tropical zones exhibited maxima in the shallow (82.99%) and deep (78.05%) layers. Boreal zones showed the strongest dependence in the middle soil layer (76.25%). Furthermore, driven by differences in transpiration rates and water retention capacity, woody plants demonstrated higher average bidirectional dependence (79.4%) than herbaceous plants (75.75%), whereas shrubland vegetation exhibited relatively lower dependence (74%). Considering the spatial distribution of climate and vegetation types, tropical zones (dominated by evergreen broadleaf forests) and boreal zones (dominated by shrublands) both showed peak bidirectional dependence in the shallow soil layer. Other climate zones, primarily influenced by grassland vegetation, exhibited peak dependence within the surface soil layer. This study reveals the variability in bidirectional dependence between vegetation productivity and multilayer soil moisture across different climatic zones and vegetation types, which will provide a robust theoretical foundation for improving regional soil moisture regulation and guiding ecosystem restoration under changing climatic conditions.