The pesticide trapping efficiency of a modified backwater wetland amended with a mixture of three pesticides, atrazine, S-metolachlor, and fipronil, using a simulated runoff event, was examined. The 700-m long, 25-m wide wetland, located along the Coldwater River in Tunica County, Mississippi, USA, was modified for hydrologic control via weirs at both ends. A pesticide mixture was amended into the wetland at the upstream weir simulating a 1-h, 1·27-cm-rainfall event from a 16 ha agricultural field. Water samples (1 l) were collected hourly within the first 24 h and again on days 2, 5, 7, 15, 21, 28 and 56, post-injection at both ends of the wetland for pesticide analysis. Peak pesticide concentrations were observed upstream 1 h after injection. Rapid pesticide removal from upstream water occurred with 63, 51 and 61% decrease of atrazine, S-metolachlor and fipronil, respectively, by 24 h. By day 7, 79, 80 and 87% decreases from peak concentrations occurred. After day 28, all pesticide concentrations were < 0·3 µg l−1, and after day 56, no target pesticides were detected. Downstream, atrazine occurred in trace amounts (<0·4 µg l−1) within 24 h and after day 28 was not detectable. S-Metolachlor occurred once downstream on day 21 (0·249 µg l−1), and fipronil was detected on days 15–56 in trace amounts (<0·05 µg l−1). Results indicate that modified backwater wetlands can efficiently trap pesticides in runoff from agricultural fields during small to moderate rainfall events, mitigating impacts to receiving waters in the main river channel. Published in 2009 by John Wiley & Sons, Ltd.