Pasture as an exclusive feeding source accounts for almost 90% of the beef consumed in Brazil and for most of the nearly 35 billion liters of milk produced annually in the country. It is estimated today that, in Brazil, grasses of the genus Brachiaria are cultivated in approximately 84 million hectares. In Brazil, the most cultivated species of Brachiaria are B. decumbens, B. brizantha, B. humidicola and B. ruziziensis. With the increasing agriculture, livestock and forestry integration, the demand for B. ruziziensis seeds has increased. It is the only fully sexual and diploid species of Brachiaria cultivated in Brazil, making it possible to carry out crosses and generate variability for further work on genetic improvement. This project had the general objective of obtaining improved cultivars of Brachiaria ruziziensis with high production of dry matter and forage quality for use in crop-livestock integration systems(both as animal feed and straw). Efforts were directed at evaluating the foraging potential and estimating genetic parameters of B. ruziziensis clones, carrying out physiological, morphological and molecular characterization of the studied germplasm, verifying the genetic variability within the species for tolerance to biotic and abiotic stresses, and carrying out intercrosses within the worked population to obtain an improved cultivar (in equilibrium). From seeds of the population originating from the first cycle of recurrent intrapopulation selection, 100 clones were selected to be evaluated in field, laboratory and greenhouse experiments. In general, the results evidenced the existence of great genetic variability among the materials, for all the evaluated traits. Many clones showed forage yields superior to the main Brachiaria cultivars in Brazil. In addition, the possibility of identification and selection of materials with superior quality of the forage produced was evidenced, increasing the quality of the species, which is already considered good. Also, for the evaluations of tolerance or resistance to the main biotic and abiotic stresses, clones with superior or similar behavior to the best experimental control were identified for all the measured traits. Even for pasture spittlebugs, whose susceptibility is the main deficiency of this species, clones with nymph survival as low as the cultivar Marandu were identified, indicating that it is possible to select B. ruziziensis materials resistant to that insect. During the project period, two intercrossing cycles of the improved population obtained in the first breeding cycle, which was the basis for the activities of this work, were carried out. Thus, the population is currently stabilized and will be subjected to tests of value of cultivation and use to confirm its productive potential and accreditation to be released as a new cultivar. All objectives of the project were achieved. Due to the great variability present within the species and the potential demonstrated by some clones evaluated, it is concluded that the Brachiaria ruziziensis breeding program should continue and has great possibilities for obtaining and distributing improved cultivars. These, in turn, will allow an increase in the productivity of beef and milk, due to the higher quality of the forage of this species in relation to the other Brachiaria species. There will also be benefits for the environment due to the greater susceptibility of the species to herbicides used in integrated systems, requiring, therefore, lower dosages of chemical products. In addition, this breeding program will provide more forage alternatives to farmers, reducing the risks of pasture vulnerability due to the cultivation of few genetic materials in large areas.