This study assessed the performance of three bioretention cells during the third and fourth years post establishment with respect to their ability to capture nitrogen (N) and phosphorus (P) in runoff from a dairy farm production area. The effects of two treatments across the three cells were evaluated: a vegetation treatment using switchgrass (Panicum virgatum L.) and a soil amendment treatment using low‐P compost (derived from leaf litter). Cell 1 has neither vegetation nor compost; Cell 2 includes vegetation without compost; Cell 3 includes both vegetation and compost. The system was installed in 2016; performance was monitored in 2018 and 2019, after vegetation was well established. In 2019, bioretention cell hydrology was modified to create an internal storage zone (ISZ) and increase hydraulic retention time (HRT), targeting improved nitrate removal. In 2018, all three cells reduced effluent concentrations of total N by >50% and of both total P and soluble reactive P (SRP) by >90%. Similar trends were found in 2019 with the ISZ, except SRP effluent concentrations were significantly higher compared with 2018, indicating a tradeoff of P leaching associated with increased HRT. Averaging eight monitored storms, median mass removals of all nutrients for Cell 2 (with vegetation and without compost) was >94%. System performance improved during the third and fourth years of operation compared with results of the initial monitoring, highlighting the importance of monitoring once plant and soil media have become established.