Five years of CDC light trap data from Suffolk County, NY, were analyzed to compare the applicability of human population density (HPD) and land use/cover (LUC) classification systems to describe mosquito abundance and to determine whether certain mosquito species of medical importance tend to be more common in urban (defined by HPD) or residential (defined by LUC) areas. Eleven study sites were categorized as urban or rural using U.S. Census Bureau data and by LUC types using geographic information systems (GISs). Abundance and percent composition of nine mosquito taxa, all known or potential vectors of arboviruses, were analyzed to determine spatial patterns. By HPD definitions, three mosquito species, Aedes canadensis (Theobald), Coquillettidia perturbans (Walker), and Culiseta melanura (Coquillett), differed significantly between habitat types, with higher abundance and percent composition in rural areas. Abundance and percent composition of these three species also increased with freshwater wetland, natural vegetation areas, or a combination when using LUC definitions. Additionally, two species, Ae. canadensis and Cs. melanura, were negatively affected by increased residential area. One species, Aedes vexans (Meigen), had higher percent composition in urban areas. Two medically important taxa, Culex spp. and Aedes triseriatus (Say), were proportionally more prevalent in residential areas by LUC classification, as was Aedes trivittatus (Coquillett). Although HPD classification was readily available and had some predictive value, LUC classification resulted in higher spatial resolution and better ability to develop location specific predictive models.