Phosphorus (P) loss from urban areas has been identified as a major contributor to declining surface water quality. The objective of this study was to determine the relationship between extractable soil P, depth of soil sampling, and dissolved reactive P (DP) concentration in runoff from turfgrass areas. At each site, runoff was generated on turfgrass and adjoining areas where turfgrass cover was removed. Across all six locations and the wide range of nutrient management schemes, variation of extractable soil P concentration and saturation ratios of 0-2cm samples accounted for 49-59% (r 2 = 0.49-0.59, n = 92) of variation of DP concentration in runoff from bare soil and soil with turfgrass cover. Despite a high degree of soil P stratification, changing sampling depth generally did not improve the relationship between soil test P and runoff DP concentrations. Across the narrower range of soil P levels common to lawns in New York (0-50mg kg −1 Morgan extractable soil P), none of the soil tests or P saturation levels (for 0-2cm depth) could accurately predict runoff P concentrations from soil with turfgrass cover (r 2 = 0.02 to 0.23, n = 72). For bare soil plots, restricting the analysis to the same range (<50mg kg −1 Morgan extractable P) did not alter the relationship between soil test P and runoff DP concentrations observed for the entire range (0-658mg kg −1 ) of soil-test P concentrations. These results suggest soil testing will not be an effective tool to predict runoff from turfgrass areas across the range of soil P levels common to New York State.