In recent years, more and more attention is focused on the characteristic changes of weather in Beijing, the frequency of fog and haze, and the changes of visibility. In virtue of regression analysis and MannKendall rank statistics, the trends and the abrupt variations of visibility are analyzed by using the data of manual air horizontal visibility at surface meteorological stations from 1980 to 2014 in the southwestern Beijing. The annual mean values of visibility show decreasing trends at the stations of plain and mountain areas. The visibility is better in spring and winter, and the diurnal variation of visibility is significantly. The analysis results show that abrupt variations of the annual mean visibility in the plain areas appeared in 1997 There existed significant correlation between visibility and surface meteorological conditions and air pollutants, with the correlation between the humidity and visibility being the best. Using air pollution data published by the local authorities in 2013, the analysis of visibility and air pollutant concentration illustrates that the correlation between PM25 and visibility is the best, while that between SO2 and visibility is the poorest.