Biomass models are key tools for estimating carbon in agroecosystems. This research was conducted in Caquetá, Colombia, in plantations and agroforestry systems. A total of 41 trees of Hevea brasiliensis and 40 of Theobroma grandiflorum were sampled to estimate the above-ground biomass (A b ), and 19 and 12 trees, respectively, were excavated to estimate the below-ground biomass (B b ). The models were adjusted based on the determination coefficient (R2), the adjusted R2, and Akaike’s and the Bayesian information criteria. The recommended models for A b in H. brasiliensis and T. grandiflorum were Ln(A b )=-2.99+2.72*Ln(DBH) and Ln(A b )=-2.59+2.48*Ln(D30), respectively (A b : kg.tree-1, DBH: diameter at breast height in cm, D30: trunk diameter at a 30 cm height). Additionally, models based on diameter and height were developed, as well as others based on the crown area. These models constitute an advancement in improving estimations of biomass and carbon, thus reaching Tier 2 in research and mitigation projects.