Abstract Mastitis, the inflammation of the mammary gland most commonly due to bacterial infection, is an important welfare concern in meat-breed sheep. The objective of our study was to evaluate udder surface temperature obtained from infrared thermography as a diagnostic tool for subclinical mastitis. We predicted that udder halves with subclinical mastitis would have greater surface temperatures than healthy udder halves. We enrolled 39 meat-breed ewes (10 Hampshire, 24 Polypay, 5 Targhee; 2 to 9 yr of age) at the University of Wisconsin-Madison sheep unit. All procedures were approved by the UW-Madison Institutional Animal Care and Use Committee. Milk samples from each udder half were collected in duplicate weekly beginning 2 to 4 d postpartum until 8 wk postpartum. Immediately before milk sampling, the udder was photographed from the rear using a thermal camera. A sample was considered a true infection if at least 100 CFU/mL of a bacterial species were isolated. From each photo, we obtained values for minimum, maximum, and mean skin surface temperature for the right and left udders. Three linear mixed models were fit using minimum, maximum, and mean temperatures as the response variables. Fixed effects were infection status (infected or not), weeks postpartum (wk 1-8), and ambient temperature, with a random effect of udder half nested within animal ID. Contrary to our predictions, infection status did not predict udder skin surface temperature (P > 0.12). Healthy udder halves had a mean temperature of 35.5 ± 0.1°C (mean ± SE), minimum temperature of 29.9 ± 0.3°C, and maximum temperature of 37.8 ± 0.1°C. Infected udder halves had a mean temperature of 35.4 ± 0.1°C, minimum temperature of 30.5 ± 0.3°C, and maximum temperature of 37.7 ± 0.1°C. Our results contradict findings in dairy cows, and suggest more research is needed to evaluate the relationship between udder surface temperature and subclinical mastitis in sheep.