The North Dakotan Turtle Mountains are an island of primarily forested habitat home to the American marten, Martes americana, a meso-carnivore in the mustelid, or weasel, family. American marten populations disappeared around 1940, but recently reappeared in this region; however, both their distribution and the effects of habitat fragmentation on their distribution are unknown. Historically, American martens have been located in the Turtle Mountains; yet current descriptions of favorable marten habitat do not match any North Dakotan habitats. I used track plates and camera traps to determine the presence/absence of martens. I determined that American martens were present. To model probabilities of marten presence/absence in the Turtle Mountains, I used these data in conjunction with landscape metrics such as amount of water, developed land, and agriculture, as well as various indices of forest fragmentation. This isolated landscape in North Dakota allowed me to ultimately verify American marten range expansion. Concerning habitat fragmentation, the way the forest patches are distributed appears to be more important to marten habitat than interior forest area, although the latter is important as well. Water is a significant predictor (p-values less than 0.05) of martens at both local and landscape scales, whereas developed land is significant (p-values less than 0.05) only at larger scales. This research will allow local and state policy makers to make informed decisions about the management of areas vital to the survival of the American marten.