Background and Objectives: There is no general consensus as to the normal limit or recommended procedure for visual fixation index (VFI) during caloric test because fixation suppression (FS) could be influenced by different variables. In this study, we present our mathematical modeling according to different variables to establish a clinical implication of VFI in patients with vertigo. Materials and Methods: Data were analyzed from a retrospective record of patients who underwent caloric testing. Three subgroups were classified (normal caloric response, unilateral vestibular hypofunction and central vertigo). VFI distribution and mathematical modeling to estimate an associated probability were performed from each group. A receiver operation characteristics (ROC) curve was plotted to determine its diagnostic value. Results: Four hundred eighteen, 67 and 14 patients met our inclusion criteria for normal caloric response, unilateral vestibular hypofunction and cerebellar infarction respectively. VFIs on warm irrigation showed more consistent distribution than cold irrigation in normal caloric response group. In contrast, there was significant inconsistency of VFIs between each side in unilateral vestibular hypofunction group (p>0.05). FS ability was inversely propositional with increase in age on all caloric (p<0.05). The area under the ROC curve of VFIs on warm caloric irrigation was 0.821 and that on cold irrigation was 0.785 for detecting central vertigo. Conclusion: Calibrated normal limit for VFI according to age is needed to ensure its clinical implication. Diagnostic value of VFI and its reliability on warm stimulation was superior to that of cold, which indicates VFI on warm stimulation seems to be a more reliable parameter.