American mink (Neovison vison) are an ecologically damaging invasive species where they have been introduced in Europe. Effectiveness of mink population control by trapping has been difficult to assess, without knowing how efficiently mink are caught by traps or detected by other methods. Use of track-recording rafts to detect mink and guide trapping effort has proved efficient and leads to a supposition that no detection indicates absence of mink. To draw this conclusion with any confidence requires a measure of detectability. We applied occupancy models to data from an earlier study to estimate detectability of individual American mink on track-recording rafts. Estimated detectability of individual mink, per raft, and 2-week check period varied between 0.4 in late summer and 0.6 in late autumn. By inference, risk of failing to detect a mink that was present would be <5% given 4–6 independent opportunities to detect it. These opportunities could be created either by using a raft spacing that ensured multiple detections of each mink or by monitoring rafts through a succession of check intervals. Within certain simple constraints, raft location did not contribute substantially to detection probability. These findings will allow field operators, strategists, and funders to assess with confidence the success of efforts to control mink density. We expect the estimation of individual detectability to be similarly valuable in population control or eradication of other species.