Phosphorus transfer from agricultural lands to water bodies via surface runoff and subsurface drainage continues to be a big environmental water quality issue. However, no model is capable to provide reasonable P losses in surface runoff and subsurface drainage. Since the EPIC model has open source code and easy to scale up, it had been applied to simulate crop yields, surface runoff and P losses in surface runoff with continuous P draw-down under corn-soybean rotation in a clay loam soil; as the first step before modification of P losses in subsurface drainage. Different potential evapotranspiration equations and curve number equations had been compared to determinate the most suitable equations (Penman-Monteith equation and curve number with soil moisture index) for this region. Simulated crop yields were almost equal to observed data with minor biases (NSE of 1.0, PBIAS of -1.3 and RSR of 0.1). Simulation of surface runoff was satisfactory with NSE of 0.5, PBIAS of -21.2 and RSR of 0.7. Simulated surface runoff flow was higher (+22.7%) than observed data due to over estimation under exceptional precipitation, which was resulted from slow vertical transport of water without considering preferential flow in EPIC. This also contributed to overestimation (+33%) of P losses in surface runoff; another possible reason was its constant P adsorption/desorption rate factors. Generally, these results demonstrate the potential of the EPIC model for adoption in the Brookston clay loam soil of the Great Lakes region with future incorporation of preferential flow and modification of P adsorption/desorption process.