Abstract: This study investigated characteristics of air quality and environmental meteorological conditions during heating period in Handan City of Hebei Province using local environmental and meteorological data, and statistically forecasted mass concentrations of six pollutants with regression and back propagation (BP) neural network methods. The results show that the air quality index (AQI) of PM_2.5, PM₁₀, SO₂, CO (95) and NO₂ are highest in winter and lowest in summer, as opposed to the O₃-8(90) AQI. During heating period, local primary pollutants are PM_2.5 and PM₁₀, and the AQI of abovementioned pollutants are higher than the annual average except for O₃-8(90). Compared with no-heating period, during heating period precipitation amount and surface air temperature are lower, weak wind occurs more frequently, and local inversion temperature and static stability index are obviously higher. The worst environmental conditions occur in January and the worse time is early morning, especially in 5-7 am. The mass concentrations of primary pollutants are positively correlated with their individual mass concentrations, inversion temperature, dew point temperature and mixing layer height in the previous day, but negatively correlated with air temperature, wind speed, visibility and mixed layer height. The results further show that the skill on forecast of pollutant mass concentration with the BP neural network method is better than those with the linear regression method, and the BP neural network method can be applied to the air quality prediction in Handan City.