The FLOTROP dataset contains numerous plant observations (around 340,000 occurrences) in northern tropical Africa (from the 5 th to 25th parallel north) in open ecosystems (savannah and steppe). They were collected by multiple collectors between 1920 and 2012 and were managed by Philippe Daget. These observations are probably the most important and unique source of plant distribution over the Sahel area. The data are now available in the Global Biodiversity Information Facility (GBIF), Tela Botanica website, and as maps in the African Plant Database. For the overall area involved, this dataset has increased by 40% the data available in the GBIF. For some countries between the 15th and 21st parallel north, the FLOTROP dataset has increased available occurrences 10-fold compared to the data existing in the GBIF. Tropical northern Africa (herein defined as between the 5th and 25th parallel north) is mostly occupied by open ecosystems, such as steppe and savannah. The vegetation in these ecosystems is consumed by animals, either wildlife or livestock, and is also used by the local communities for food, energy or medicinal purposes. The open ecosystems in tropical northern Africa are of great importance to the economy, food security and human well-being. Plant diversity within these ecosystems is driven by many factors, such as the climate, soil, fire and grazing. Plant diversity in these regions is being greatly impacted by global change. Historical data are needed to understand species and diversity dynamics. The database presented in this work is the collection of numerous datasets gathered over the years. At the outset, the FLOTROP database was intended to store all the data recorded by IEMVT (French institute for tropical livestock production and veterinary medicine, now part of CIRAD) in the sixties. In 1993, CIRAD and CNRS set up a project to collect a maximum of botanical surveys within these regions. Two software packages were created by the team to manage the database. The first was created under DOS then a second was started under Windows using the APL DYALOG language. Data were collected and scanned between 1993 and 2016. We extracted the data from the software version. We shared the species occurrences recorded in the database on the Tela Botanica website (http://www.tela-botanica.org/) and on the GBIF database.