Abstract Forest carbon (C) estimates are the key inputs to the understanding of the global C cycle. We report the estimates of forest carbon pool and its spatial distribution in the Indian forests for the years 1994 and 2010 at 5 km grid level. This study improves upon earlier spatial estimates of Indian forest biomass carbon by using data from a robustly designed National Forest Inventory (NFI). The realized sampling intensity has addressed the large heterogeneity of the Indian forest types and allowed the computation of 5 km grid level forest C, yielding a realistic estimate of forest biomass C in Indian forests. Forest cover density maps were intersected with 5 km mesh and estimates of forest area, forest carbon density for each Agro-ecological sub region and forest carbon pools were linked to the 5 km grid coverage of India. National forest carbon estimates for the years 1994 and 2010 are 3911.78 and 4368.03 TgC respectively, and these estimates showed a net increase of 456.25 TgC in 16 years. Uncertainty of the estimates has been addressed spatially. Mean forest carbon density increased from 61.14 Mg ha−1 in 1994 to 64.08 Mg ha−1 in 2010. C densities for dense and open forest in 1994 estimated as 77.08 and 38.47 Mg ha−1 with total C pools of 2895.28 TgC and 1016.50 TgC which has increased to 80.24 Mg ha−1 and 41.69 Mg ha−1 with total C pools of 3176.48 TgC and 1191.55 TgC in 2010. This study provides the first 5 km level C analysis for Indian forests. Spatial distribution of C shows large differences in C density over Indian forests indicating that estimates of the spatial distribution of C are even more important than the total C pool estimates of the country.