N deficiency is one of the major factors affecting pastures yield. This work aimed to evaluate the decomposition of signal grass (Brachiaria decumbens) and gliricidia (Gliricidia sepium) or “sabiá” (Mimosa caesalpiniifolia) biomass, separately and mixed in different proportions, on N supply to soil.  Two experiments were established with the litter bag technique each with six treatments (100% unfertilized signal grass, 100% N fertilized signal grass, 75:25, 50:50, 25:75 signal grass:gliricidia or “sabiá” ratios, and 100% gliricidia or “sabiá”) and nine incubation times (0, 2, 4, 8, 16, 32, 64, 128 and 256 days). At each sampling time, determinations were conducted for total organic matter, and organic carbon, total nitrogen and lignin contents, and C:N and Lignin:N ratios were calculated. The signal grass and gliricidia mixtures increased decomposition rates, compared to the single treatments. The highest N liberation rates for signal grass:gliricidia mixtures were found for 25:75 signal grass:gliricidia, with k values of 0,0254 g.g-1.day^-1, while for signal grass:”sabiá” the fastest decomposition was for the 75:25 mixture, releasing  0,0064 g.g^-1.dia^-1. While mixed grass:legumes litter C:N ratios did not noticeably change over time, the pure grasses litter had important reductions in the C:N ratios. For both mixtures, lignin content was increased with the mixture of legume material into the litter, and as expected, litter content increased over time, indicating that this inclusion may help to increase soil organic carbon over time, since litter is an important component of soil humus. At the same time, the inclusion of legumes increased nitrogen release from the litter, although the effect was stronger for gliricidia then for “sabiá”. There is a need though to combine the evaluation of the decomposition per se with litter production to evaluate the possible effects of inclusion of shrub-tree legumes in nutrient recycling in signal grass pastures.