Dry eye syndrome is a common irritating eye disease. Current clinical diagnostic methods are invasive and uncomfortable to patients. A custom designed noncontact infrared (IR) thermal image system was developed to measure the spatial and temporal variation of the ocular surface temperature over a 6-second eye-opening period. We defined two parameters: the temperature difference value and the compactness value to represent the degree of the temperature change and irregularity of the temperature distribution on the tear film. By using these two parameters, in this study, a linear discrimination result for the dry eye and the normal eye groups; the sensitivity is 0.9, the specificity is 0.86 and the receiver operating characteristic (ROC) area is 0.91. The result suggests that the custom designed IR thermal image system may be used as an effective tool for noncontact detection of dry eye.