<p>Coffee agroforestry is increasingly promoted as a nature-based solution for climate change mitigation, yet spatially explicit information on carbon stocks in smallholder systems remains scarce, particularly in heterogeneous landscapes such as Kalimantan. This study quantified the spatial characteristics and carbon stock potential of robusta coffee agroforestry in Malinau Regency, Kalimantan, by integrating field-based inventories with multisource remote sensing. Thirty 20 × 20&#xa0;m plots were established in four villages and used to measure aboveground biomass, belowground biomass, litter, and soil carbon. High-resolution UAV orthomosaics were combined with Sentinel-2 and NICFI–Landsat imagery to capture canopy heterogeneity and delineate plot boundaries, while vegetation indices were used as predictors in land-cover classification. Robusta coffee agroforestry systems showed highly diverse spatial configurations, ranging from clustered and mixed mosaics to more uniform patterns. Mean total carbon stock reached 103.48&#xa0;Mg C ha⁻¹, ranging from 63.44 to 157.79&#xa0;Mg C ha⁻¹ across plots. Soil carbon was the dominant pool, with an average of 92.25&#xa0;Mg C ha⁻¹ and a contribution of approximately 89.15% to total carbon stock, whereas aboveground biomass, belowground biomass, and litter contributed 3.58, 0.89, and 6.76&#xa0;Mg C ha⁻¹, respectively. The integration of UAV and satellite imagery effectively captured fine-scale canopy variation that is often missed by medium-resolution products alone. These findings indicate that soil carbon is the dominant carbon pool within the sampled robusta coffee agroforestry systems and provide a plot-based basis for carbon stock assessment in tropical coffee agroforestry landscapes. The combined field and remote sensing approach can support the further development of evidence-based monitoring, reporting, and verification frameworks.Query</p>