Visibility impairment has become an important environmental issue receiving great attention from both the scientific community and the public. Long-term meteorological data from Baoji, China, were collected to investigate the trend in visibility change from 1980 to 2012. The 33-year average visual range is 12.0 km. The best 20% of the visibility observations in a calendar year shows a general decreasing trend from 1994 onwards, while the worst 20% exhibits a slight increasing trend from 1997 onwards. These results suggest the progressive degradation of air quality in Baoji in recent years. Intensive PM2.5 measurements were conducted from March 2012 to February 2013 to determine the causes of visibility impairment. An analysis based on IMPROVE equation reveals that PM2.5 organic matter (OM) contributes to 34.2% of the light extinction coefficient (b(ext)) on an annual basis, followed by (NH4)(2)SO4 (30.0%), NH4NO3 (20.1%), elemental carbon (9.2%) and soil dust (6.5%). The largest contributor to hex, for the Worst 20% group is (NH4)(2)SO4, and the contribution of NH4NO3 for the Worst 20% group increases by a factor of similar to 3 compared with the Best 20% group. Source apportionment using a positive matrix factorization receptor model indicates that secondary sulfate is the main source of PM2.5 (23.0%), followed by fugitive dust (20.5%), coal combustion (19.9%), secondary nitrate (15.5%), biomass burning (14.3%) and motor vehicle emissions (6.8%). These quantitative results could be useful for policy makers to take effective measures to control the haze pollution in Baoji. Further, the results also are likely to be relevant for other mid-sized cities in China.