Phosphorus (P) contributes to eutrophication of surface waters and buffer strips may be implemented to reduce its transfer from agricultural sources to watercourses. This study was conducted to test the hypothesis that soil type and slope influence the retention of dissolved organic P and inorganic orthophosphate in agricultural runoff in 2-m-wide buffer strip soils. A solution, comprised of dissolved orthophosphate and the organic P compounds glucose-1-phosphate, RNA, and inositol hexakisphosphate (1.8 mg L−1 total P) and a chloride tracer, was applied as simulated overland flow to grassland soil blocks (2 m long × 0.5 m wide × 0.35 m deep), containing intact clay or loam soils, at slope angles of 2, 5, and 10°. Phosphorus forms were determined in the surface and subsurface flow from the soil blocks. Slope had no significant effect on the hydrological behavior of the soil blocks or on the retention of any form of P at the water application rate tested. The clay soil retained 60% of the unreactive P and 21% of the reactive P applied. The loam soil retained 74% of the unreactive P applied but was a net source of reactive P (the load increased by 61%). This indicates leaching of native soil P or hydrolysis of organic compounds and complicates our understanding of P retention in buffer strip soils. Our results suggest that a 2-m buffer strip may be more effective for reducing dissolved unreactive P transfers to surface waters than for reducing the eutrophication risk posed by dissolved reactive P.