Surveillance of pathogen richness in wildlife is needed to identify host species with high zoonotic spillover risk. Many predictors of pathogen richness in wildlife hosts have been proposed, but these predictors have mostly been examined separately and not at the pantropical level. Here we analyzed 15 proposed predictors of pathogen richness using a model ensemble composed of bagged random forests, boosted regression trees, and zero-inflated negative binomial mixed-effects models to identify predictors of pathogen richness in wild tropical mammal species. After controlling for research effort, species geographic range area was identified to be the most important predictor by the model ensemble while the most important anthropogenic factor was hunting pressure. Both fast-lived and slow-lived species had greater pathogen richness, showing a non-linear relationship between the species fast-slow continuum of life history traits and pathogen richness, whereby pathogen richness increases near the extremities. The top species with the highest pathogen richness predicted by our model ensemble are Vulpes vulpes, Mus musculus, Canis lupus, Mustela erminea, and Lutra lutra. Our results can help support evidence-informed pathogen surveillance and disease reservoir management to prevent the emergence of future zoonotic diseases.