We explored different treatments to enhance the probability of sowed seeds of two early successional (ES, Cecropia obtusifolia and Ochroma pyramidale) and two late successional (LS, Brosimum costaricanum and Dialium guianense) species to escape predation and germinate in abandoned cattle-raising pasture fields in Southeastern Mexico. ES species were sown in groups of 50 seeds under three treatments: invertebrate exclusion, burial, and exposition to seedeaters. LS species were sown in groups of 10 seeds under three treatments: vertebrate exclusion, burial, and exposition to seedeaters. We registered seed predation and germination 2, 4, 8, 16, 32, and 64 days after the initial sowing. Overall, ES showed higher predation rates (mean ± SE = 0.45 ± 0.07 seed seed−1 day−1; n = 3) than LS species (0.09 ± 0.02 seed seed−1 day−1). Cecropia obtusifolia was completely predated in all treatments after 8 days. Burial and exclusion treatments reduced final predation in circa 6% for O. pyramidale, relative to that of exposed seeds (85% after 8 days); most germination occurred in buried seeds (3.7%). In B. costaricanum, burial enabled germination by 10%; exposed and excluded seeds were removed 100%. Dialium guianense showed 12% germination in buried seeds and circa 20% of the seeds were not removed after 64 days. Direct sowing would be a recommended rainforest restoration practice for species with relatively large seeds if deposited in groups and buried. Studies which address variation across numerous sites are necessary in order to generate more consistent seed predation patterns and rainforest restoration principles in tropical pastures.