Four sets of eight or twelve 0.7 ha plots, designed for soil and hydraulic uniformity, were used to screen potential Best Management Practices (BMPs) for reducing total phosphorus (TP) concentrations and loadings in the Everglades Agricultural Area (EAA) of south Florida. The four production systems and their alternatives (treatments) studied were: (1) sugarcane (interspecific hybrids of Saccharum sp.) versus drained fallow plots; (2) fast versus slow drainage rates for sugarcane; (3) rice (Oriza sativa L.) in rotation following radishes to serve as a P filter crop versus traditional flooding fallow; and (4) banding phosphorus (P) fertilizer at 50% of the soil-test recommendation rate for cabbage (Brassica oleracea L.) versus full-rate broadcast applications. The study showed that there were no differences in P concentrations in drainage water between sugarcane and drained fallow fields. Annual P loading to the plots in rainfall and irrigation water (0.63 kg TP ha1) exceeded the P loading of drainage waters (0.52 kg TP ha1 for sugarcane and 0.59 kg TP ha1 for drained fallow plots). Slow drained sugarcane plots exhibited significantly higher TP concentrations than the fast drained plots. However, TP loads were significantly higher (0.97 kg ha1) for fast drained plots than for the slow drained plots (0.67 kg ha1). Rice as a P filter crop following radishes reduced TP concentrations and loadings. Finally, banding P fertilizer at a reduced rate for cabbage reduced TP concentrations compared to those for broadcasted P at the full recommended level. Total phosphorus loadings in drainage water were 1.17 kg ha1 for banded and 1.38 kg ha1 for broadcast treatments. A total of 1.30 kg TP ha1 entered the plots in rainfall and irrigation water. All treatment TP loadings leaving the plots in drainage water were close in magnitude to TP loadings to the plots, even under heavy fertilization. This indicates that the EAA system is currently a net assimilator of P.