Traditional weed control practices in Mexico use legumes as cover crops or manures. Legumes used in these practices play a dual role in agroecosystems by protecting the soil from erosion and by enriching it with organic matter and N through Rhizobium symbiosis. Farmers in the tropical regions of Mexico use Mucuna spp., Canavalia spp. and other legumes to control weeds in their fields. We conducted in vitro bioassays and greenhouse experiments to evaluate the toxic effect of four legumes velvetbean [Mucuna deeringiana (Bort) Merr.], jackbean [Canavalia ensiformis (L.) DC.], jumbiebean [Leucaena leucocephala (Lam.) de Wit], and wild tamarind [Lysiloma latisiliquum (L.) Benth.] on weed growth, and on the survival of insects and nematodes. The aqueous leachates (1%) of the four legumes were tested on three test plants by seed germination and radicle growth bioassays in petri dishes. The aqueous leachates of all four legumes exhibited strong phytotoxic effect on the radicle growth of the test plants. The effects of velvetbean and jackbean leachates were also evaluated on the survival of phytopathogenic nematodes. Both leachates had nematotoxic effects. Greenhouse experiments were conducted to evaluate the effect of the four legumes dried leaves incorporated as mulches to potting soil, on the number and biomass of weeds, and on the biomass of tomato (Lycopersicon esculentum L.) plants. The decomposition of velvetbean and jackbean leaves in potting soil significantly reduced (>50%) the development of phytopathogenic nematodes in the roots of tomato. A 5‐yr field experiment conducted at the University of Yucatan (UADY) evaluated the effect of velvetbean and jackbean used as living cover crops, and jumbiebean and wild tamarind used as dead mulches incorporated on soil surface, on weed growth, and corn (Zea mays L.) yield. The experimental field was treated with the traditional slash and burn system in February 1994. In July 1994 the experiment was performed using the local agricultural practices in a complete randomized block design with three repetitions. The treatments were: corn+velvetbean, corn+jackbean, corn+jumbiebean, corn+wild tamarind, corn+Paraquat (1,1′‐dimethyl‐4,4′‐bipyridinium ion) herbicide, corn weeded by hand, and plots without corn weeded by hand. The number, biomass, diversity, and relative importance of weeds, as well as corn yield, were evaluated. In addition, taxonomic composition of weeds was determined. All legumes reduced weed growth with velvetbean (as living cover crop) producing the largest weed biomass reduction (68%). These legumes also improved the yield of corn during the first 2 yr of the experiment. For better management of natural resources, the use of legumes as biological tools in agriculture to control weeds and improve soil conditions should be encouraged through coordinated efforts between farmers, academic, and governmental institutions.