Extreme temperature is strongly associated with human health, but few studies are available for the effects of extreme temperatures on respiratory disease (RD) in rural villages in Jiuquan, China over 2018-2019. In this study, we evaluate the effect of daily mean temperature on RD hospitalizations in suburban rural villages. A distribution lag non-linear model was constructed to analyze the relationship between the temperature and the daily risk of RD hospitalizations; and stratified analysis by sex and age group was performed. The effect of daily mean temperature on RD hospitalizations was non-linear and lagging. With a reference 25th percentile of temperature (-1.8 °C), exposure to extremely low temperature (-15.8 °C, 1st percentile), or moderate low temperature (-8.3 °C, 10th percentile) were associated with 1.396 (95% CI: 1.251, 1.558) and 1.216 (95% CI: 1.180, 1.253) increased risks of RD hospitalizations over lag 0-21 days. For RD morbidity, the effects of moderate high temperature (90th percentile of temperature, 22.5 °C) appeared at on the exposed day (lag 0), with the largest hot effect (RR 1.008, 95% CI 1.001, 1.020), while the effects of extremely high temperature (99th percentile of temperature, 27.0 °C) were insignificant. The effect of ambient temperature varied with gender and age. Both cold and high temperatures have more serious influence on males than females. In contrast, the elderly (age ≥ 65) seemed to be more sensitive in extremely low temperature (RR = 3.471; 95% CI, 2.183, 5.518; lag 0-21). The effect of moderate high temperature on the < 65 years group appeared for the first time at lag 0 (RR = 1.012; 95% CI, 1.001, 1.029). Both high temperature and low temperature increased the RD hospitalization risk; the harmful effect and duration of low temperature were greater than that of high temperature; the ≥ 65-year group and male were more sensitive to low temperature.