Based on the visibility, temperature, pressure, humidity, wind and atmospheric particulate concentration data in Chongqing, the characteristics and influencing factors of visibility in Chongqing are analyzed. The visibility forecast model is established using the neural network method, and the effects of introducing the PM_2.5 concentration factor on the forecast model are analyzed and compared. The results show that the visibility distribution in Chongqing is low in the west, high in the east, and low along the Yangtze River. The average visibility during fog is lower than that during precipitation and much lower after removing the data of rain and fog days, indicating that low visibility is more affected by water vapour in the atmosphere. The proportion of fog increases significantly in winter, leading to a significant decrease in average visibility in winter. The precipitation increase in June and October is an important reason for the significant decrease in average visibility in these two months. The diurnal variation of visibility shows a single-peak pattern, and the periods of high fog and precipitation overlap with the areas of low average visibility, which is an important reason for the low visibility at night. Atmospheric humidity, temperature and particle concentration are all critical factors affecting visibility. When relative humidity is less than 70%, visibility decreases significantly with the increase of PM_2.5, and when relative humidity is more than 70%, the influence of PM_2.5 on visibility decreases. The prediction effect of introducing the PM_2.5 concentration factor into the visibility objective forecast model is better than that of not introducing the PM_2.5 concentration factor, especially the prediction effect in autumn and winter is significantly improved.