The ecosystem services (ES) concept is meant to facilitate consideration of the value of nature in conservation and landscape management processes by translating ecosystem functions into human benefits. Incorporating the ES concept into policy and decision-making has proven difficult due to challenges in identifying, measuring, and locating services and in predicting the impacts of decisions upon them. ES mapping offers a key solution to increase our understanding of the spatial patterns of ES supply and demand and the spatial relationships between them, but may be challenging to implement given a lack of spatial data related to ES or existence of such data at coarse resolution that may not facilitate accurate ES quantification, mapping and modeling. This issue is particularly acute in urban settings where landscapes are highly heterogeneous and fragmented. This research seeks to improve our understanding of urban ES supply, demand and the relationships between them, as well as the impacts of spatial scale, input data quality and method choice on ES mapping in urban landscapes. The dissertation is composed of three studies. In the first study, I introduce a spatially-explicit framework for quantifying and mapping ES supply and demand using carbon storage and sequestration services as an example. This framework assesses supply based on biophysical conditions and demand based on socioeconomic characteristics, allowing for more integrative ES assessments in urban areas. In the second study, I evaluate the sensitivity of ES maps to input spatial data resolution and method choice (ecosystem component-based and land-cover proxy-based methods) in a heterogeneous urban landscape using biomass carbon storage as an example. I find that ES map accuracy is highly dependent on analytical scales and input data representativeness. ES estimates based on ecosystem-component data are more accurate than those based on land-cover proxies. The accuracy of land-cover proxy-based maps, however, can be increased by using high-resolution land-cover maps. The third study aims to increase understanding of ES supply, demand, and supply-demand balance in urban contexts. To this end, I create a high-thematic-resolution land-cover dataset and combine it with the InVEST pollination model to assess the capacity of urban ecosystems to supply pollination services to satisfy the demands of urban agriculture. I find using land-cover dataset at a higher thematic resolution enhances the accuracy of pollination estimates, highlighting the importance of considering scale and land-use dependencies in urban ES mapping. Combined, these studies advance our knowledge of ES supply, demand and the relationships between them, and provide new insight into the impacts of input data spatial and thematic resolution and method choice on the accuracy of urban ES maps.