Implementing Mechanisms for Reducing Emissions from Deforestation and Forest Degradation (REDD +) requires knowledge of the carbon storage potential in different forest types and different key reservoirs. We investigated the relationship between tree diversity and carbon stock in two forest types in Cameroon, an evergreen forest (EGF) in the South and a semi-deciduous forest (SDF) in the East. We aimed to characterize the two forest types, evaluate carbon stocks in five forest carbon reservoirs, and assess the relationship between the carbon stocks of the three pools (living carbon, dead carbon and soil carbon) studied and the diversity of the trees. An inventory of tree species with dbh ≥5 cm was made, covering 84 quadrats (25 m × 25 m), along 7 transects. Total woody biomass and dead wood were determined by allometric equations. We found that the EGF was more diverse in tree composition than the SDF. Absolute density and basal area were significantly higher in SDF than in EGF. There was a significant difference between living carbon of 327.35 t C.ha−1 in EGF and 369.77 t C.ha−1 in SDF. The contribution of dead wood in carbon storage varied on average from 8.40 t C.ha−1 in the EGF to 11.19 t C.ha−1 in the SDF. Similarly, trees of 5 cm ≤ dbh up to 10 cm dbh varied on average from 6.17 t C.ha−1 and 5.02 tC.ha−1 respectively for EGF and SDF. There was a positive but weak correlation between total carbon and tree species richness and a positive and moderate correlation between total carbon and biological types in both forest types. The results of this study highlight the need to take into account the assessment of carbon stocks of all reservoirs, and types of forests for a successful implementation of REDD + in the Congo Basin.