Urban landscapes are increasingly recognized as providing important ecosystem services (ES) to their occupants. Yet, urban ES assessments often ignore the complex spatial heterogeneity and land-use history of cities. Soil-based services may be particularly susceptible to land-use legacy effects. We studied indicators of three soil-based ES, carbon storage, water quality regulation, and runoff regulation, in a historically agricultural urban landscape and asked (1) How do ES indicators vary with contemporary land cover and time since development? (2) Do ES indicators vary primarily among land-cover classes, within land-cover classes, or within sites? (3) What is the relative contribution of urban land-cover classes to potential citywide ES provision? We measured biophysical indicators (soil carbon [C], available phosphorus [P], and saturated hydraulic conductivity [K ]) in 100 sites across five land-cover classes, spanning an ~125-year gradient of time since development within each land-cover class. Potential for ES provision was substantial in urban green spaces, including developed land. Runoff regulation services (high K ) were highest in forests; water quality regulation (low P) was highest in open spaces and grasslands; and open spaces and developed land (e.g., residential yards) had the highest C storage. In developed land covers, both C and P increased with time since development, indicating effects of historical land-use on contemporary ES and trade-offs between two important ES. Among-site differences accounted for a high proportion of variance in soil properties in forests, grasslands, and open space, while residential areas had high within-site variability, underscoring the leverage city residents have to improve urban ES provision. Developed land covers contributed most ES supply at the citywide scale, even after accounting for potential impacts of impervious surfaces. Considering the full mosaic of urban green space and its history is needed to estimate the kinds and magnitude of ES provided in cities, and to augment regional ES assessments that often ignore or underestimate urban ES supply.