The Agricultural Policy Environmental Extender (APEX) model is used to simulate the effects of vegetative filter strips on runoff and pollutant loadings from agricultural watersheds.  A long-term paired watershed study under corn (Zea mays L-soybean [Glycine max (L.) Merr.] rotation with agroforestry (grass+trees), and grass upland buffers (4.5 m width), in Northeast Missouri, has shown 11-35% measured reduction in sediment and nutrient loads.  The objective of this study was to calibrate and validate the APEX model for the study watersheds and find optimum buffer dimensions and placement locations.  ArcAPEX and APEX0604 versions were used for the simulations.  The simulated corn and soybean yields were within ±13% and ±27% of the measured yields, respectively.  The agroforestry, grass buffer, and control watershed models were calibrated (1998 to 2001) and validated (2002 to 2008) for daily event-based runoff with r² and Nash-Sutcliffe Coefficients (NSC) values of 0.7-0.8 and 0.6-0.7, respectively.  The event-based total phosphorous (TP) loads were calibrated and validated well with r² and NSC ranging 0.7-0.8 and 0.6-0.7, respectively.  The sediments and total nitrogen were not calibrated.  The simulated grass and agroforestry buffers reduced 16% and 17% average event-based runoff and 15% and 50% TP, respectively.  An increase in buffer widths to 5.5 and 7.5 m did not effectively change runoff or TP loads.  Placement of buffers in backslope positions was little more effective for reducing runoff and TP loads in the agroforestry watershed.  Simulation of a winter cover crop reduced TP loadings by 12% to 19%.  The calibrated and validated APEX model may be used to  simulate various management scenarios to determine environmental benefits of buffers.