Core areas are important descriptors of animal space-use patterns, but current estimation methods rely on arbitrary rules and potentially lead to imprecise or erroneous area estimates. We proposed a Bayesian statistical model that incorporates an individual-based method for estimating core area boundaries. The model accounts for boundary uncertainty and multiple scales of clustering by partitioning a home range into ≥2 completely spatially random point patterns defined by a kernel density isopleth. We used data from coyotes (Canis latrans), bobcats (Lynx rufus), and red-shouldered hawks (Buteo lineatus) to estimate core areas for individual animals. We also estimated core areas from simulated point patterns with known boundaries, varying numbers of points, and relative densities of points inside core areas, and compared estimates to those obtained using the 50% isopleth. Optimal isopleths for the empirical data ranged between 18.7% and 71.5%. We found no species-specific range of core area isopleths. Across all simulated scenarios, our method outperformed the 50% isopleth-based estimate, which consistently overestimated core areas. Minta overlap values were 20–40% higher across all scenarios for our method compared to the 50% isopleth. Minta overlap values were >75% in 90% of scenarios using our method. Objectively estimating core areas using our individual-based method may lead to improved inference about which behavioral and ecological processes underlie observed space-use patterns because of greater estimate precision.