To analyze human thermal environments in protected horticultural houses (plastic houses), human thermal sensations estimated using measured microclimatic data (air temperature, humidity, wind speed, and solar and terrestrial radiation) were compared between an outdoor area and two indoor plastic houses, a polyethylene (PE) house and a polycarbonate (PC) house. Measurements were carried out during the daytime in autumn, a transient season that exhibits human thermal environments ranging from neutral to very hot. The mean air temperature and absolute humidity of the houses were 14.6-16.8℃ (max. 22. 3℃) and 7.0-12.0 g∙m-3 higher than those of the outdoor area, respectively. Solar (K) and terrestrial (L) radiation were compared directionally from the sky hemisphere (↓) and the ground hemisphere (↑). The mean K↓ and K↑ values for the houses were respectively 232.5-367.8 W∙m-2 and 44.9-55.7 W∙m-2 lower than those in the outdoor area; the mean L↓ and L↑ values were respectively 150.4-182.3 W∙m-2 and 30.5-33.9 W∙m-2 higher than those in the outdoor area. Thus, L was revealed to be more influential on the greenhouse effect in the houses than K. Consequently, mean radiant temperature in the houses was higher than the outdoor area during the daytime from 10:45 to 14:15. As a result, mean human thermal sensation values in the PMV, PET, and UTCI of the houses were respectively 3.2-3.4℃ (max. 4.7℃), 15.2-16.4℃ (max. 23.7℃) and 13.6-15.4℃ (max. 22.3℃) higher than those in the outdoor area. The heat stress levels that were influenced by human thermal sensation were much higher in the houses (between hot and very hot) than in the outdoor (between neutral and warm). Further, the microclimatic component that most affected the human thermal sensation in the houses was air temperature that was primarily influenced by L↓. Therefore, workers in the plastic houses could experience strong heat stresses, equal to hot or higher, when air temperature rose over 22℃ on clear autumn days.