AbstractPacific lampreys Entosphenus tridentatus (formerly Lampetra tridentata) are declining in the Columbia River basin, and the use of large, main‐stem river habitats by larvae of this species is unknown. We used a deepwater electrofisher to explore occupancy, detection, and habitat use of larval Pacific lampreys and larval Lampetra spp. in the lower Willamette River, Oregon. We used a generalized random tessellation stratified approach to select sampling quadrats (30 × 30 m) in a random, spatially balanced order. Pacific lampreys, Lampetra spp., and unidentified lampreys were found in the Willamette River; larvae were detected in all areas except the Multnomah Channel. We calculated reach‐ and quadrat‐specific detection probabilities and the amount of sampling effort required for 80% confidence that larval lampreys were in fact absent when they were not detected. Lampreys were detected in a variety of areas (although relatively low numbers were collected), including shallow, nearshore areas; midchannel areas (depth up to 16 m); and anthropogenically affected areas. Detection probabilities (i.e., in occupied areas) were 0.07 (reach) and 0.23 (quadrat). The sampling effort required for 80% confidence that lampreys were absent when undetected was 20 quadrats (in the lower Willamette River reach) and 6 subquadrats (within a quadrat). Differences in lamprey detection by depth were not observed. A variety of sizes was collected (20–144 mm total length), indicating the likely occurrence of multiple ages of larvae. Our study identifies how the occurrence of larval Pacific lampreys can be quantified with statistical rigor in a large river (i.e., larger than fourth order [1:100,000 scale]). The effect of channel management activities on larval lampreys should be considered in efforts to conserve these important species.