Where carnivore species exist at low densities, are cryptic, and inhabit forested habitats where detection is low, survey approaches commonly rely on observation of tracks. Recent advances in probability sampling for aerial surveys of track networks in snow show promise for improving density estimates, but they are not applicable when continuously following a track network is not possible. Occupancy–abundance modeling is an alternative survey approach for wide-ranging carnivores, but past models may not be appropriate when using tracks because tracks of one individual or group of individuals may extend across several survey units. We derived an occupancy–abundance relationship by simulating the intersection of wolf (Canis lupus) travel paths and survey grid cells for a range of wolf-pack densities. Wolf movement paths were simulated using a habitat-biased, correlated random walk movement model using step lengths and turning angles of 17 Global Positioning System (GPS)-collared wolves in west-central Alberta, Canada. We estimated occupancy levels for a range of pack densities found in North America and found that pack density was linearly related to the proportion of occupied survey units. Concurrent movement paths of 5 GPS-collared wolf packs were used to evaluate the model predictions. While the model overestimated the number of packs by 33%, the difference translated into only 13 wolves across the study area. We discuss improvements for continued development of occupancy surveys as a potential method to determine carnivore abundance where other approaches are not feasible. © 2012 The Wildlife Society.