This paper presents an uncertainty analysis of subjective thermal comfort measurement, which was overlooked in some subjective thermal comfort studies. Three chamber experiments were conducted to collect the data and to analyze how the uncertainty of subjective measurement is influenced by the survey instrument and the experiment protocol. Four widely-utilized subjective indicators were examined: thermal sensation, humidity sensation, draught sensation and thermal satisfaction. It was found that the Standard Uncertainty, defined as the Standard Error of the Mean, of subjective thermal comfort measurement would decrease with the increasing number of person-votes (number of subjects times votes per subject). When the person-votes were above 40, the Standard Uncertainty of subjective measurement was within 3% of the total scale length, confirming the precision and reliability of subjective thermal comfort measurement. It was found the measurement of thermal sensation/satisfaction and humidity sensation had similar uncertainty level, while the measurement of draught sensation had higher uncertainty. Discrete multiple-points scales (Likert) are recommended for subjective thermal comfort measurement since continuous scales enlarge intra-individual differences.