Abstract: Based on the monthly maximum temperature and minimum temperature data in Yuncheng City from 1961 to 2020, the annual extreme values of maximum temperature and minimum temperature in different return periods were fitted and calculated by Pearson type-Ⅲ, Gumbel and Log-normal distribution methods.Three measures of fitting standard deviation, fitting relative deviation and Kolmogorov test were also used to give a comparative superiority evaluation.The results show that for nearly 60 years, both the mean maximum (minimum) temperature and extreme maximum (minimum) temperature in Yuncheng City show an increasing trend, and the increasing trend of the low temperature is significantly greater than that of the high temperature; the annual average extreme maximum temperature and minimum temperature are mainly concentrated from June to July and late December to early February, respectively.The best lognormal distribution fitting effects are identified in annual extreme maximum temperature and extreme minimum temperature in Yuncheng City, while the extreme maximum (minimum) temperature fitted by Gumbel distribution tends to be higher (lower) than the observed data.