Accurate determination of sex in harvested species is critical for understanding demography and developing population models for management. We used genetic-based sex identification to assess accuracy of external carcass and pelt examination at registration and maximum canine root area (MRA) to determine sex of American martens (Martes americana) trapped in the Upper Peninsula of Michigan, 2000–2004. Overall percent similarity between MRA and genetic-based sex determination was 98.4% (n = 188). In contrast, only 84.6% (n = 421) of martens were similarly classified using external examination. For external examination, percent similarity to genetic-based sex determination for juveniles (<1-yr old) and adults (≥1-yr old; Wald χ²₁ = 2.168, P = 0.141), as well as for males and females (Wald χ²₁ = 0.005, P = 0.946), was similar. We recommend MRA as a suitable technique for sex determination of martens; thus, marten sex and age (using cementum annuli counts) can be obtained from one lower canine tooth. We do not recommend use of external examination at registration to identify sex of martens without implementing additional quality assurance measures.