While attention to logging in the tropics has been increasing, studies on the long-term effects of silviculture on forest dynamics and ecology remain scare and spatially limited. Indeed, most of our knowledge on tropical forests arise from studies carried out in undisturbed tropical forests. This bias is problematic given that logged and disturbed tropical forests are covering now a larger area than the so-called primary forests. A new network of permanent sample plots in logged forests, the Tropical managed Forests Observatory (TmFO), aims to fill thisgap by providing unprecedented opportunities to examine long-term data on the resilience of logged tropical forests at regional and global scales. This presentation aims to introduce the TmFO which currently includes 24 experimental sites distributed across three tropical regions, with a total of 490 permanent plots and 921 ha of forest inventories. We will present here the main objectives and research questions of TmFO as well as its methodological approach and preliminary results. (Texte intégral)