Tropical trees constitute an ecologically important functional group in terrestrial ecosystems because of the essential roles that they play in sustaining biodiversity and carbon storage. The persistence and evolutionary potentials of tropical trees are, however, increasingly threatened by human-induced rapid changes in abiotic and biotic environments. For long-lived forest trees, gene dispersal by seeds and pollen is critical for tracking shifting climatic niches and for maintaining genetic variation needed to adapt to changing environments. Understanding the potential responses of tropical trees to environmental changes depends in part upon quantifying the rates of seed and pollen dispersal. This dissertation aims to quantify the spatial extent and magnitude of seed and pollen dispersal and their respective genetic impacts in a comparative context, by focusing on four Neotropical tree species that have distinct dispersal and pollination syndromes and life-history strategies. By using parentage inference and inverse modeling, I found that long-distance gene dispersal by seeds is common in these vertebrate-dispersed tropical trees, in which models predicted 1–18% of dispersal events exceeding 1 km. This fraction of pollen dispersal >1 km could reach 10–20% in these species. Furthermore, simulations with gene dispersal distances realistically represented suggest that seed and pollen dispersal limitation can lead to genetic diversity loss in tropical tree populations. By examining the respective genetic impacts of seed vs. pollen dispersal, I found that seed dispersal is the primary force driving spatial genetic patterns in these species. It suggests that the functional loss of seed-dispersing vertebrates, as a result of anthropogenic disturbance in tropical forests, could alter not only tree population spatial structure and ecological dynamics, but also genetic structure and evolutionary dynamics.