Competition for light in mixed-species plantations is a main limiting factor for tree growth. Understanding the light absorption and light use efficiency for each species at different planting densities and spatial arrangement of trees is essential for improving these plantations management. The study was conducted in Itatinga-SP, Brazil. A complete randomized block design was set up with 4 blocks and I O treatments per block with pure and mixed plantations of Eucalyptus grandis (E) and Acacia mangium (A), a N2-fixing species (NFS). The treatments were composed with different planting densities (6 m x 3 m, 3 m x 3 m and 3 m x 2 m) and arrangements (I 00% A, I 00% E, 50% A and 50% E, 33% A and 67% E) between species. Tree growth and biomass were monitored at 38, 45 and 52 months after planting. The absorbed photosynthetically activeradiation (APAR) for each tree was simulated with the MAESTRA model. The leaf area of each tree, leaf angle distributions and leaf area density was estimated in situ to parameterize the model. Hemispherical photos were taken over the same period of time (every 2 months) and used to test MAESTRA simulations and validate the light interception simulations. Computed Light Use Efficiency (LUE) for trunk wood production was estimated as the ratio of the wood growth and the MAESTRA simulated APAR, showing contrasted patterns of LUE for each treatment. The results will give insights to choose the best design, decreasing light competition and improving the association between eucalyptus and NFS, for a more sustainable management of pure and mixed forest plantations.