Conflicting findings have emerged regarding the presence of attentional biases (ABs) in health anxiety, probably due to methodological limitations in the stimuli used in cognitive tasks and the assessment of health anxiety-relevant factors. The current study sought to examine ABs toward health-related threats using idiographically chosen health-threat words in a non-clinical sample. A modified dot-probe task using idiographically selected health-threat words was administered to an undergraduate sample. Self-report measures were administered to assess somatic, cognitive, and behavioral aspects of health anxiety, in addition to assessing negative affect, anxiety sensitivity, and experience of actual medical conditions. Results showed that behavioral and somatic aspects of health anxiety were significantly associated with AB toward personally relevant threat words, even after controlling for negative affect, anxiety sensitivity, and experience of actual medical conditions. Additional analyses revealed that these biases reflected difficulty disengaging attention from threat rather than a facilitated detection of threat. In contrast, illness-related cognitions were found to be unrelated to ABs. These findings suggest an association between threat-related ABs and excessive health-care seeking efforts.