Livestock grazing is common and widespread throughout North America, yet few studies have evaluated its effects on small mammals. We studied small mammals in mixed-conifer forests and oak woodlands on the Cascade–Siskiyou National Monument in southern Oregon, USA, to 1) evaluate small-mammal microhabitat associations, 2) identify riparian-associated species, and 3) test the hypothesis that grazing does not influence small mammals after accounting for microhabitat associations. We live-trapped small mammals at 16 study sites and used logistic regression to model probability of capture on measured habitat characteristics at each trap station and to evaluate effects of grazing. Over 2 years, we trapped 1,270 individual small mammals representing 18 species. Odds of capturing western harvest mice (Reithrodontomys megalotis), dusky-footed woodrats (Neotoma fuscipes), and long-tailed voles (Microtus longicaudus) were lower (P < 0.05) on heavily versus lightly grazed sites. Odds of capture for deer mice (Peromyscus maniculatus) were lower (P < 0.05) on heavily versus lightly grazed sites in woodlands, but there was less difference in the odds of capture between grazing intensities in conifer forests. Odds of capturing Townsend's vole (Microtus townsendii) were lower on heavily versus lightly grazed riparian areas. Western harvest mice, long-tailed voles, and Townsend's voles were associated with, but not obligated to, riparian areas. Deer mice were ubiquitous, but captures were also higher (P < 0.05) in riparian areas compared with uplands. Siskiyou chipmunks (Tamias siskiyou) and piñon mice (Peromyscus truei) were associated with uplands (P < 0.05) rather than riparian areas. Trowbridge's shrews (Sorex trowbridgii), Siskiyou chipmunks, and bushy-tailed woodrats (Neotoma cinerea) were positively associated with coarse woody debris. Land managers should anticipate that small mammals associated with herbaceous or shrub cover, particularly in riparian areas, will decline when cattle remove this cover.