In Central Africa, forest planning relies on a method of forest production simulation that lacks precision and is liable to compromise logging sustainability. The aim of this study was to improve growth forecasts for four of the main commercial species in Central Africa. A survey was set up in south-eastern Cameroon, covering 387 trees: 136 Assamela (Pericopsis elata), 54 Moabi (Baillonella toxisperma), 124 Sapelli (Entandrophragma cylindricum) and 73 Tali (Erythrophleum suaveolens). Their diameter was measured regularly from 2009 to 2012. Competition intensity was quantified on the scale of a circular plot with a radius of 20m around each target tree. Topographic and hydrological characteristics were estimated with the aid of a geographic information system. Growth was modelled on the basis of tree diameter and growth conditions (competition index and local environment). The tree growth figures currently used in Cameroon are identical to those observed in forests undisturbed by logging in the case of Assamela, Moabi and Tali, but they overestimate the growth of Sapelli by 20%. A log-normal model incorporating tree diameter and competition intensity produced growth estimations for the four species with a determination coefficient (R2) ranging from 0.092 for Assamela to 0.47 for Moabi. The topographic and hydrological variables did not significantly improve the quality of the forecasts. These models may be used to forecast growth on the basis of data from forest inventories as currently drawn up for planning purposes in Central Africa, in order to improve the definition of management parameters for the four species studied, such as minimum logging diameters or regeneration rates.