Tropical forests in Africa store large amounts of carbon and there is an urgent need for accurate methods to estimate this important carbon stock. One method to achieve this is by allometric equations but in many forest ecosystems in Africa these have not been developed. This study combined biomass data of 896 trees from five tropical countries in Africa and eight different sources to develop mixed-species regression equations for estimation of total biomass and height in Dry, Moist and Wet forest types. For estimation of total biomass, allometric equations combining diameter, height and wood density provided the best estimators in the three forest types. Because adding wood density to diameter improved height estimation, we recommend using allometric equations that combine diameter and wood density for height estimations in mixed and diverse tropical forests. Comparing ecosystem-specific (Dry, Moist and Wet) allometric equations to general allometric equations developed with combined data, and also to pan-tropical equations, showed that ecosystem-specific equations provided better estimators. The results highlight the importance of considering wood density in tree allometry for biomass as well as for tree height estimations. Although general allometric equations can be useful, this study recommends when they are available, the use of existing site-specific or ecosystem-specific allometric equations which provide better estimates.