Abstract: High quality data series are not only the base of our study on climate change, but also the base of the diagnosing, forecasting and valuating climate. Because of sites moving and observation method changing, the climate data have been inhomogeneity for a long time. The original data applied in this paper sourced from Beijing meteorological information center, and were homogenized by the method of MASH (Multiple Analysis of Series for Homogenization). The MASH package consists of an interative procedure designed to detect possible break points through mutual comparisons of a number of climate data series, in which each series is not necessarily homogeneous. Possible break points and shifts were detected and adjusted through mutual comparisons among all available series. It can be seen from historical records that, since 1960, the site of Beijing meteorological observatory has been moved for five times, and the moving was taken place in 1965, 1969, 1970, 1981 and 1997 respectively. This paper describes annual mean yearly temperature time series before and after homogenized correction. It can be found that the difference between them is obvious. The discontinuity points in the series are in 1964, 1980 and 1996 respectively. As can be seen that they are all near the year when the site was moved. It illuminates that it is necessary to homogenize the original data. Based on the homogenized daily maximum and minimum temperature in meteorological stations of Beijing during 1960-2009, the temporal characteristics of extreme temperature variation and the response of extreme temperature to urbanization during 1960-2009 were studied. The results show that, in Beijing, the warming rate of annual average temperature was 0.39℃ per 10 a in recent 50 a, and 0.22℃ per 10a higher than the national average level. But the warming rate was unsymmetrical. The warming rate of annual average minimum temperature was much higher than that of annual average maximum temperature over the same period. That means that, in recent 50 years, climate warming in Beijing is mainly manifested as the rising of lowest temperature. As a whole, the extreme high temperature and extreme low temperature varied significantly with the rate of linear increasing tendency of 0.23℃ per 10a and 1.02℃ per 10 years, and the abrupt change of climate warming happened about 1988. In 1997-2009, Beijing had the most high temperature days and the least frost days in the past 50 a. From 1980 to 2009, the extreme high temperature and the high temperature days increased more in urban areas, and less in suburbs and exurbs of Beijing. The extreme low temperature and the low temperature days decreased more in urban areas and suburbs, and less in exurbs. Results show that the variations of extreme temperature have been obvious affected by the process of urbanization in Beijing. This result will reference on studying fundamentals about urbanization effect on urban climate.