Emergence and spread of antimicrobial resistance (AMR) are major concerns for human and animal health worldwide. Consequently, there is increasing pressure to reduce antimicrobial use (AMU) in food-producing animals. In this project, we studied the association between AMU and AMR in non-aureus staphylococci (NAS) isolated from Canadian dairy herds. Distinct types of dry cow therapy (DCT), antimicrobials, and routes of administration were compared. The AMR profile was determined using a micro-broth dilution method against a panel of 23 antimicrobials for 1,702 NAS isolates obtained from 89 herds. A subset of these isolates was submitted to whole-genome sequencing and presence of AMR genes was determined using data from 4 databases. Antimicrobial use was determined for all herds using an inventory of empty drug containers and quantified for each antimicrobial as number of antimicrobial daily doses (ADD) administered. Data on management practices, including type of DCT, were obtained using questionnaires. Only systemic AMU was associated with prevalence of AMR in NAS, whereas intramammary or intrauterine use were not. Three drug classes, all of high or very high importance for human medicine, were associated with drug-specific AMR when administered systemically: penicillins, 3rd generation cephalosporins, and macrolides. Prevalence of tet, erm and bla_ARL genes in NAS was higher in herds that used more tetracyclines, macrolides, and 3rd generation cephalosporins, respectively. Mean AMU of antimicrobials administered as DCT was lower in herds using selective DCT in comparison to herds using blanket DCT. However, use of selective DCT was not associated with prevalence of AMR in NAS. The either weak or non-existent association between AMR and antimicrobials administered intramammarily suggest that a decrease in AMR of NAS following implementation of selective DCT would be minimal in comparison to reduced use of systemic antimicrobials