Long-term monitoring data indicate a persistent decline in American Kestrel populations across North America. Loss or alteration of habitat have been listed as potential causal factors, but basic information on kestrel space use, including breeding home range size, is lacking. No study has provided robust estimates of the ranging behavior of breeding kestrels based on tracking data of any resolution. We fitted 19 adult female kestrels with solar-powered GPS transmitters during the incubation period in northern Virginia. High-resolution tracking began during the early nestling stage for 17 birds. We collected an average of 1710 locations per bird through the end of the breeding season (31 August), with 13 birds tracked through the fledging of their young. Autocorrelated kernel density home range estimation showed that female kestrels used breeding home ranges that were smaller (average: 0.32 km2) than most previously published range sizes. Home ranges did not vary significantly in size across breeding stages and demonstrated little overlap with the ranges of neighboring kestrels. Five females shifted their territories in the post-breeding stage (i.e., after disappearance or dispersal of fledglings) between 1.5 and 12.3 km from their nest box; they maintained these new ranges at least to the migration period. We also documented home range excursion forays (n = 128) by all 12 consistently tracked females. Mean (4.0 km) and maximum (127.7 km) foray distances were some of the largest reported among birds and mammals relative to home range size. Weekly foray rates were highest during the nestling stage, and for birds that ultimately shifted from their breeding home range. The existence of long-distance foray behavior and the use of multiple summer home ranges, both shown here for the first time for this species, has a direct impact on interpretation of kestrel nest-site and habitat selection data, and on the assessment of potential threats to this species in the breeding season.