Appropriate quantification of the spatial distribution of lead (Pb) in urban soils requires an understanding of the influence that individual landscape features have on soil Pb concentrations as well as the importance of the larger landscape context. In order to examine patterns of residential soil Pb concentrations we employed a sampling scheme that addressed both landscape context and individual landscape features predicted to affect Pb retention. The larger landscape context was considered by examining the distance to major road networks and housing age. Individual landscape features that were hypothesized to affect lead retention included trees, lawns, buildings, and roads. We intensively sampled 61 residential lots in Baltimore City using field portable x-ray fluorescence. Lead concentrations exceeded the USEPA reportable limit of 400 ppm at 53% of sites. Field measurements ranged from 7 to 9151 ppm with a mean value of 363 ppm and a median value of 124 ppm. Lead contamination was concentrated close to buildings and major road networks. All data were used as input to a general liner model (GLM) and the resulting parameter estimates were used in ArcGIS 9.3 to map the spatial distribution of Pb in these soils. The resulting empirically based model predicts the spatial distribution of Pb in soil as a function of landscape features, including distance to major road networks, distance to buildings, housing age and significant interaction variables. The model increases our understanding of the importance of both landscape context and individual landscape features to the spatial distribution of a persistent environmental pollutant.