A pilot-scale, ponded wetland consisting of an open pond and a vegetated pond in series was constructed on a cotton farm in northern New South Wales, Australia, and assessed for its potential to remove pesticides from irrigation tailwater. Ten incubation periods ranging from 7 to 13 days each were conducted over two cotton growing seasons to monitor removal of residues of four pesticides applied to the crop. Residue reductions ranging 22-53% and 32-90% were observed in the first and second seasons respectively. Average half-lives during this first season were calculated as 21.3 days for diuron, 25.4 days for fluometuron and 26.4 days for aldicarb over the entire wetland. During the second season of monitoring, pesticide half-lives were significantly reduced, with fluometuron exhibiting a half-life of 13.8 days, aldicarb 6.2 days and endosulfan 7.5 days in the open pond. Further significant reductions were observed in the vegetated pond and also following an algal bloom in the open pond, as a result of which aldicarb and endosulfan were no longer quantifiable. Partitioning onto sediment was found to be a considerable sink for the insecticide endosulfan. These results demonstrate that macrophytes and algae can reduce the persistence of pesticides in on-farm water and provide some data for modelling.