AbstractManagement practices are available that may reduce sediment or phosphorus (P) loading to streams draining agricultural land. An evaluation of the long term effects of implementing one or more of these practices is needed. We used the Erosion Productivity Impact Calculator (EPIC) and Agricultural Non‐Point Source (AGNPS) models to evaluate 20 management practices for sediment and P yields from 4 ha fields in a 1272 ha watershed of dairy farms. Five crop sequences [continuous corn (Zea mays L.) CC(OA)AAA, CC(OA)AAAA, CCC(OA)AAA, and CCCAAA; 0, oat (Avena sativa L.); A, alfalfa (medicago sativa L.)] combined with either fall moldboard plowing (mFMP), chisel plow (mFCP), no‐till (mNT), or no‐till substituting anhydrous ammonia for manure (aNT), were imposed on a toposequence of three Alfisols. The normal sequence, 2 yr of corn, oat/alfalfa seeding, 3 yr of alfalfa with fall application of solid manure followed by moldboard plowing, CC(OA)AAA‐mFMP, served as reference. Scenarios were subjected to 42 yr of simulated local climate. Scenarios with 43% or less row crop and mNT on a 4% sloping soil reduced sediment yield from 24 to 47%, at edge of field, but increased total P yield from 158 to 338%, mostly as soluble P. The EPIC output parameters (soil loss ratio [C] for June, erodibility [K], Manning's n, and SCS curve number [CN]) served as inputs to the (AGNPS) model, a cell based single rainfall event model that simulates runoff, sediment, and P yields from a watershed. Crop sequences not exceeding 43% corn and mFCP or aNT reduced watershed sediment yield 13 to 41%, but increased total P yield by 44 to 113% as compared with CC(OA)AAA‐mFMP. Average annual costs were $11 ha−1 less for these sequences when compared with CC(OA)AAA‐mFMP.