Megafauna (terrestrial vertebrate herbivores > 5 kg) can have disproportionate direct and indirect effects on forest structure, function, and biogeochemical cycles. We reviewed the literature investigating these effects on tropical forest dynamics and biogeochemical cycles in relation to ecology, paleoecology, and vegetation modelling. We highlight the limitations of field‐based studies in evaluating the long‐term consequences of loss of megafauna. These limitations are due to inherent space‐time restrictions of field‐studies and a research focus on seed dispersal services provided by large animals. We further present evidence of a research gap concerning the role of megafauna in carbon cycling in tropical ecosystems. Specifically, changes in aboveground biomass might not be noticeable in short‐term studies because of slow vegetation dynamics requiring decades to respond to disturbance (i.e. defaunation). Nutrient cycling has received even less attention in relation to the role of megafauna in tropical forests. We present an approach to investigate the effects of megafauna from new perspectives and with various tools (notably, vegetation models), which can simulate long‐term dynamics in different environmental and megafauna density scenarios. Vegetation models could facilitate interaction between plant–animal ecology and biogeochemistry research. We present practical examples on how to integrate plant–animal interactions in vegetation models to further our understanding of the role of large herbivores in tropical forests.