Identifying the main predictors of species’ extinction risk while accounting for the effects of spatial and phylogenetic structures in the data is key to preventing species loss in tropical forests through adequate conservation practices. We recorded 22 705 precise geographical locations of primate occurrence across four major geographic realms (Neotropics, mainland Africa, Madagascar and Asia) to assess predictors of threat status using a novel Bayesian spatio-phylogenetic approach. We estimated the relative contributions of fixed factors (forest amount, body mass, home range, diel activity, locomotion, evolutionary distinctiveness and climatic instability) and random factors (space and phylogeny) to primate extinction risk. Precipitation instability increased the extinction risk in the Neotropics but decreased it in mainland Africa and Madagascar. Forest amount was negatively associated with extinction risk in all realms except Madagascar. Body mass increased the extinction risk in the Neotropics and Madagascar, whereas home range increased the extinction risk in mainland Africa and decreased it in Asia. Evolutionary distinctiveness negatively influenced extinction risk only in mainland Africa. Our findings highlight the importance of climate change mitigation and forest protection strategies. Increasing the protection of large primates and reducing hunting are also essential.