We used rendezvous site locations of wolf (Canis lupus) packs recorded during 1996–2006 to build a predictive model of gray wolf rendezvous site habitat in Idaho, USA. Variables in our best model included green leaf biomass (Normalized Difference Vegetation Index), surface roughness, and profile curvature, indicating that wolves consistently used wet meadow complexes for rendezvous sites. We then used this predictive model to stratify habitat and guide survey efforts designed to document wolf pack distribution and fecundity in 4 study areas in Idaho. We detected all 15 wolf packs (32 wolf pack-yr) and 20 out of 27 (74%) litters of pups by surveying <11% of the total study area. In addition, we were able to obtain detailed observations on wolf packs (e.g., hair and scat samples) once we located their rendezvous sites. Given an expected decrease in the ability of managers to maintain radiocollar contact with all of the wolf packs in the northern Rocky Mountains, rendezvous sites predicted by our model can be the starting point and foundation for targeted sampling and future wolf population monitoring surveys.