An ecological niche modelling (ENM) approach was used to predict the potential feeding and spawning habitats of small (5-25kg, only feeding) and large (> 25kg) Atlantic bluefin tuna (ABFT), Thunnus thynnus, in the Mediterranean Sea, the North Atlantic and the Gulf of Mexico. The ENM was built bridging knowledge on ecological traits of ABFT (e.g. temperature tolerance, mobility, feeding and spawning strategy) with patterns of selected environmental variables (chlorophyll-a fronts and concentration, sea surface current and temperature, sea surface height anomaly) that were identified using an extensive set of precisely geo-located presence data. The results highlight a wider temperature tolerance for larger fish allowing them to feed in the northern – high chlorophyll levels – latitudes up to the Norwegian Sea in the eastern Atlantic and to the Gulf of Saint Lawrence in the western basin. Permanent suitable feeding habitat for small ABFT was predicted to be mostly located in temperate latitudes in the North Atlantic and in the Mediterranean Sea, as well as in subtropical waters off north-west Africa, while summer potential habitat in the Gulf of Mexico was found to be unsuitable for both small and large ABFTs. Potential spawning grounds were found to occur in the Gulf of Mexico from March-April in the south-east to April-May in the north, while favourable conditions evolve in the Mediterranean Sea from mid-May in the eastern to mid-July in the western basin. Other secondary potential spawning grounds not supported by observations were predicted in the Azores area and off Morocco to Senegal during July and August when extrapolating the model settings from the Gulf of Mexico into the North Atlantic. The presence of large ABFT off Florida and the Bahamas in spring was not explained by the model as is, however the environmental variables other than the sea surface height anomaly appeared to be favourable for spawning in part of this area. Defining key spatial and temporal habitats should further help in building spatially-explicit stock assessment models, thus improving the spatial management of bluefin tuna fisheries.