ABSTRACT World wet tropical forests, and especially the ones in the Colombian Pacific area, are the target of a small tree (minor diameter) selective harvest process, used in short-cycle industries, such as bioenergy. This situation generates a reduction in stored carbon and biomass, and becomes an emission of greenhouse gases (GHG). Allometric models for aboveground biomass are few, despite being an important tool of carbon calculation. The goal of this study was to develop multi-species allometric models for small trees aboveground biomass in wet tropical forests. A total of 61 individuals (diameter at breast height -DBH- < 12 cm) was measured, cut and weighed to estimate their biomass. The model with the best adjustment was selected considering criteria of determination coefficient (R2) and adjusted R2, mean quadratic error of prediction, Akaike and Bayesian Information Criteria and the biological logic of the model. Best-fit allometric model (R2= 0,72) was with DBH and total height as independent variables, considering that it is a multi-species model coming from forests with a high diversity.