Introduction: The reliability of thermal quantitative sensory testing (QST) has yet to be fully established. In this study we investigated intra- and interrater reliability of thermal QST in a blinded manner. Methods: Two investigators recorded thermal detection and pain thresholds on the hand of 22 volunteers, twice on two occasions. Results were analyzed using descriptive statistics, intraclass correlation coefficients (ICCs), and coefficients of variation (CVs). Results: Mean intraindividual differences were small for all measures except cold pain thresholds. ICC values for intra- and interrater reliability were: cold detection, 0.27-0.55; warm detection, 0.33-0.69; and heat pain, 0.39-0.86. Cold pain yielded high ICC values (0.87-0.94), but also high CV (84.9-90.2%). Conclusions: In young, healthy adults, thermal detection and heat pain thresholds of the hand demonstrated good reliability for group comparisons and individual analyses. Cold pain threshold measures may be suitable for group comparisons, but a large variance in the data limits individual analyses.