The development of suitable heat-resistant material and reasonable thermal structure design for a hypersonic aircraft requires highly precise heat transfer prediction, especially at the wing tips or stagnation points where the thermal environment is particularly rigorous. Unfortunately, existing techniques for measuring heat flux via thermal sensors in hypersonic test facilities are excessively complex; any slight deviation from ideal conditions may lead to inaccuracy. In this study, the influence of installation on the accuracy of heat flux measurement for a sphere model using cylindrical thermal sensors was numerically investigated. The sensors examined were protruding or recessed from the model surface on the order of 0.1-0.5 mm, and the influence of variations in installation was estimated by comparing the results against the smoothly installed sensor. Experiments were also run using different Reynolds numbers and sensor diameters to ensure the rules and mechanisms reported are as comprehensive as possible. Results showed that " unsmooth" (i. e., protruding, recessed) installation creates substantial deviation from the actual heat transfer rate; protruding installation resulted in larger deviation than recessed, and the deviation of either installation configuration grew more severe at larger depths from the model surface.