Understanding the light absorption and light use efficiency of each species at the tree scale is essential to disentangle the effects of intra- and inter-species interactions on productivity in mixed-species forest plantations. A complete randomized block design was set up using Eucalyptus grandis (E) and Acacia mangium (A), which is a N2-fixing species, planted in monospecific stands (100A, 100E) and in additive (25A:100E, 50A:100E, 100A:100E) and replacement (50A:50E) mixtures. Tree size and biomass were monitored over the complete rotation (6 years). The absorbed photosynthetically active radiation (APAR) for each tree in the experiment was simulated over the full rotation with the MAESTRA model. Measurements of tree leaf area, leaf angle distributions, leaf area density, and leaf and soil optical properties were performed to parameterize this model. The APAR and the ratio of stem biomass increment divided by the APAR (which is a measure of the Light Use Efficiency [LUE] for stem production) were calculated at tree and plot scales for each year of the rotation. The LUE of the 100E stand increased with age until stabilizing at 4 years of age, while the LUE of the 100A stand decreased between 2 and 4 years of age and increased between the two last years of the rotation. Eucalyptus trees dominated Acacia trees in mixed plantations. The stratification of the canopy led to an increase of stand Leaf Area Index (LAI) and APAR compared to monospecific plantations. However, both Eucalyptus and Acacia LUE decreased at the end of the rotation in the mixed-species stands, with the decrease occurring more markedly in Acacia, and the final stem biomass of the stand was not enhanced in mixed-species plantations compared with the average of the pure stands. Our results indicate that a stratified canopy may offer the potential benefit of capturing more light in mixed-species forests, but this may be negated if another resource deficiency prevents trees from converting intercepted