Fire is a critical ecological disturbance with both ecological benefits and socio-economic implications. Across fire-prone ecosystems, different fire regimes can be found, reflecting a combination of climatic factors and of different plant species characteristics. While the role of plant characteristics such as fuel load and flammability has long been recognized, recent theoretical studies have emphasized how the composition of other plant traits linked to fire responses, but also to growth and competitive ability can be strongly related to the fire regime and to the resilience of whole ecosystems. Here we will focus on how plant traits related to flammability and fire responses, but also to growth and competitive ability, are related to the fire regime and the overall ecosystem resilience in different ecosystems and for various climates around the globe. To address this aim, we combined statistical and mathematical modelling. We expanded on an existing dynamical model representing vegetation succession and vegetation-fire feedback. We also analysed global datasets of plant traits (e.g., TRY, AusTraits, Flamits), community composition (from sPlot), climate data, and remote sensing fire data (from MODIS/VIIRS). We show that traits related to flammability are increasingly present in communities with more frequent fires, both for forests and for open ecosystems. Noticeably instead, trait related to fire responses (e.g. resprouting, bark thickness) display different relationship with fire frequency for forest and open ecosystems. The model results also show that plant trait syndromes were fundamental in determining fire regimes and the occurrence of different plant communities under the same climatic and environmental conditions. Our findings provide valuable insights for fire management, ecosystem restoration, and reforestation efforts by highlighting the complex influence of community functional composition on fire dynamics. These findings underline also the importance of including the plant fire response when modeling fire ecosystems, for example, to predict the vegetation response to invasive species or to global change.