Abstract: Based on the observation data of National Basic Weather Stations in Sichuan for 2009-2018,the temporal and spatial distribution characteristics of low visibility in Sichuan Basin were studied.The SWC model visibility forecast was corrected by using probability density function matching method based on the correction threshold that was obtained by combining the relationship between low visibility and surface meteorological elements.The analysis shows that:(1)The area with low-visibility is mainly located in the southeast of Sichuan basin,Meishan,Leshan and Eastern Chengdu.The whole low-visibility area has a obvious tendency to move northwardly , and the distribution of low-visibility days increases significantly in autumn; the distribution of low visibility days in winter is similar to that in autumn,but areas with low-visibility is concentrated in the northeast of the basin.Low visibility episodes usually start at 0:00-8:00 and ends at 7:00-12:00 during one day.(2)The low visibility < 1 km mainly appears in stations where wind speed is less than 2 m/s,the relative humidity is basically above 90%,the surface temperature is lower than 15 °C and the 24-hour pressure change is between-2 to 10 hPa.(3)The model forecasted visibility in some low-visibility areas can be corrected by using the double correction methods that combined probability density function with meteorological factors.After correction,TS scores were significantly improved.