•The Atlantic Forest is the biome with the highest number of papers on ecological restoration.•Pantanal, Pampa and Caatinga are the least studied in terms of ecological restoration.•In all biomes, most of the studies are related to monitoring.•The number of species used in restored areas is limited.•It is necessary to include different life forms, restoration techniques and studies using reference areas to measure success. The Bonn challenge aims at the restoration of 350 million hectares of degraded ladscapes by 2030. In Brazil, the restoration goal for 2030 is 12 million hectares. Despite the great demand for ecological restoration across the whole of Brazil, there have been no analyses of the studies carried out in different biomes. In addition, conservation efforts must cover all biomes, so that different regions can take advantage of the many benefits of restoration. Our aim was to identify advances and gaps in current restoration knowledge in order to guide future efforts in Brazil. Our bibliometric survey in the Web of Science using 23 keywords related to restoration generated a total of 530 papers, of which 291 were included in the analysis. The papers were published in 121 scientific journals between 1988 and 2018, with the largest number of papers in 2016. The Atlantic Forest was the biome with the highest number of studies, as it is one of the most threatened tropical forest regions in the world and maintains the largest number of research institutions and receives the highest level of funding support in the country. Regarding the types of studies, temporal monitoring was more frequent in the Amazon, Cerrado, Caatinga, and Pampa, while the monitoring at one point in time was more frequent in the Atlantic Forest. From the studies examined, 31% used a reference area for comparing restoration success. The most studied organisms were plants (81%), and among them, trees were the most frequent, followed by fungi, birds, invertebrates, mammals, and reptiles. The pre-restoration degradation differed among biomes, with deforestation for logging the most cited in the Amazon, agriculture, and livestock in the Atlantic Forest and Cerrado, logging and cattle ranching in Caatinga, and livestock in the Pampa and Pantanal. In general, active/assisted natural succession was the most frequent restoration process: planting seedlings more readily occurred in the Amazon, Atlantic Forest, and Caatinga, whereas natural regeneration in the Cerrado and Pantanal and sowing in Pampa. The studies varied among the age of restoration (>1 to 67 years for active restoration and >1 to 120 years for passive/unassisted natural succession), and the number of species planted (1 to 121 species). We identified an important regional knowledge gap for the Pantanal, Caatinga, and Pampa, as well as the need to include reference areas, evaluate different restoration techniques (besides planting seedlings), and the inclusion of other taxa and life forms in biodiversity studies apart from trees. We also identified the need to expand research to assess landscape metrics, prioritization, legislation, and public policies.