There has been a decreasing trend in visibility during the most recent 30 years in China, especially in the east. To investigate the possibility of a multivariate nonlinear regression relationship between visibility and meteorological variables, surface meteorological data (1980-2012) collected from 521 stations were analyzed in this study. The results showed that the average variance of long-term visibility variation caused by relative humidity, wind speed, and temperature is 22%. The regression equation passed the significance test (significance level:α=0.05) at 68 stations (13%), with a 55% average variance. These stations were then divided into three groups according to different meteorological factors. Stations effectively influenced by relative humidity were found to be concentrated in Southeast China, while stations effectively influenced by wind speed are spread over most areas of China, except Xinjiang, and stations where air temperature is the dominant influence are mostly located in a band across central China. Different meteorological factors show different relationships with visibility:Relative humidity and air temperature influence the linear trend of visibility, while wind speed influences the interdecadal variation. The long-term downtrend of visibility in East China is mainly dominated by the decreasing trend of relative humidity above 70%, and the interdecadal variation of wind speed below 4 m/s.