Non-contact temperature assessment is highly valuable for screening animal populations, but its performance can be influenced by various external and biological variables. This study aimed to quantify the association between eye temperature (ET) and rectal temperature (RT) in dogs, and to identify physiological and environmental factors associated with these readings. Field study: ET was obtained with a thermal camera up to 3 times per visit (~10-min intervals), followed by RT measurement. Eye temperature was defined as the highest temperature recorded from either eye. A linear mixed-effects model evaluated ET in relation to capture time (1–3), RT, age, sex, body size, coat length, ambient temperature, and relative humidity. Device temporal stability assessment: To account for potential camera drift observed in our previous research, temperature at a fixed target inside an insulated box was recorded every 5 min for 1 h and analyzed with repeated-measures mixed model regression. Results. We analyzed 129 observations from 78 dogs. In the final ET model, both RT (β = 0.77 °C; 95% C.I. 0.512–1.019 °C) and humidity (β = 0.01 °C; 95% C.I. 0.002–0.013 °C) were positively and significantly associated with ET. The device stability assessment demonstrated no significant time-related measurement drift. Overall, ET was associated with RT but was not interchangeable and ET was slightly affected by environmental humidity. These results support further investigation of using ET as a rapid, noncontact temperature monitoring procedure.