There is a growing body of evidence that habitat fragmentation resulting from anthropogenic land use can alter the transmission dynamics of infectious disease. Baylisascaris procyonis, a parasitic roundworm with the ability to cause fatal central nervous system disease in many mammals, including humans, is a zoonotic threat, and research suggests that parasite recruitment rates by intermediate hosts are highly variable among forest patches in fragmented landscapes. During 2008, we sampled 353 white-footed mice (Peromyscus leucopus) from 22 forest patches distributed throughout a fragmented agricultural ecosystem to determine the influence of landscape-level habitat attributes on infection rates of B. procyonis in mice. We characterized each mouse in terms of infection status and intensity of infection, and calculated (on a patch-wide basis) prevalence, mean abundance of B. procyonis, and mean intensity of infection. We used an information-theoretic approach to develop a suite of candidate models characterizing the influence of landscape attributes on each of our measured characteristics of B. procyonis infection in white-footed mice, based on previous knowledge of raccoon (Procyon lotor) ecology and B. procyonis distribution in agricultural ecosystems. We observed evidence of B. procyonis infection in mice across all 22 habitat patches sampled. However, parasite recruitment rates and intensity were highly variable among patches, and the results of our analyses suggest that spatial variability in B. procyonis infections was primarily driven by emergent properties of fragmented ecosystems. In particular, prevalence, abundance, and intensity of B. procyonis infections in mice were negatively associated with the size and connectivity of forest patches. These results support previous studies indicating that habitat fragmentation can alter the transmission dynamics of infectious disease, and suggest that factors below the scale of landscape, i.e., fine-scale habitat structure or demographic and behavioral attributes of intermediate and/or definitive hosts, also may be important for predicting patterns of B. procyonis infection in intermediate hosts.