Plants and soils are in constant interaction, and net effects of all physical, chemical and biological processes that contribute to these interactions are resulting in a so-called ‘plant-soil feedback’ that is known to be critical for the restoration of original plant communities and ecosystem properties. The previous land use on sites where the Atlantic Forest is to be restored has altered the historical contingency of soil properties, which will influence forest restoration by legacy effects, however, little, if anything, is known about plant-soil feedback in the Atlantic Forest. Our aim is to investigate how plant-soil feedbacks operate in primary and secondary Atlantic Forest, and how feedbacks in currently de-forested areas may influence opportunities for successful restoration of original tree species diversity and forest ecosystem functioning. First, we will collect a large-scale observational data set on belowground species richness and abundance of microbes (bacteria, fungi, protists) and small invertebrates (nematodes), to characterize soil networks. We will link these data to information on plant/tree composition and soil abiotic properties in original and restored forests, as well as deforested land to be used for forest restoration. Then, we will empirically examine plant-soil feedbacks of these soils in order to determine their suitability for forest regeneration. For this purpose, field plots that can generate spatially explicit information on a range of variables through time will be established along 100 transects across the different landscapes. By combining ecosystem models and network data from field plots, we will evaluate which soil conditions provide best possibilities for forest restoration. By generating spatial and temporal information on a range of variables through time and scale, including positive and negative interactions from the aboveground and belowground communities, we will determine if these effects are more positive in soils from original or successfully restored vegetation than in soil from degraded or unsuccessfully restored systems, which may then be related to the collected soil biodiversity information. Data on disturbed areas with less or more suitable opportunities for plant-soil feedback interactions to forest plant/tree species will allow us to advice how to improve and speed up an Atlantic Forest restoration.