High temperature is a growing threat and impacts public health through different exposure mechanisms. Our study constructs a comprehensive exposure measurement based on temperature variability, duration, and effective influence range. We investigate human responses to high temperatures through self-rated health scores based on individual-level data from China Labor-force Dynamic Survey (CLDS). Results show that higher temperature and temperature variability significantly decrease self-rated health scores. Subjective health risk is most significantly related to the cumulative temperature in the previous two weeks. We also find that the exposure effects at night and on weekdays are more severe. Workers who experience greater exposure from commuting and work environments are negatively impacted by high temperatures. In addition, men, the elderly, middle and low education groups, rural residents are more likely to be impacted by high temperatures.