Résumé : The coffee agroforests located in the District of Kodagu, India, are known to have a high tree diversity, including rare, endangered and threaten species. However, in the last decades, changes in the tree composition of the coffee agroforests have been observed; farmers are partially or completely replacing the native trees by the exotic Australian species Grevillea robusta, which decreases the biodiversity of the coffee plantations. As the causes for the retention (or not) of certain native shade trees still remains poorly understood, the present study aims to fill this gap, as well as to model the distributions of the most common species in coffee agroforests of the region under current and future climate projections. Twenty species were analysed and modelled, and questionnaires were used with farmers to evaluate their preference regarding species and preferred traits. Most of the preferred species by the farmers can be considered multifunctional, given their ability to provide several ecosystem services, from fruit production to shade provision and soil improvement. Regarding the species distribution modelling (SDM), most of the species (19 out of 20) were found to be negatively affected by climate change in the three climate change scenarios evaluated. The findings of the present research will be integrated within the Diversity for Restoration tool (https://www.diversityforrestoration.org/), which aims to help farmers and practitioners with the selection of appropriate tree species for climate-resilient tropical forest landscape restoration.