To investigate the characteristics of chemical constitute and pollution sources of aerosol fine particulate matter during haze-fog day in Beijing in winter 2013. The samples of PM2.5 were collected in Beijing from January to February, 2013. The technique of ICP-MS and ICP-AES coupled with procedure of bathing-ultrasonic extraction was applied to determine the concentration of 40 elements in the aerosol samples to analyze the characteristics of elements distribution statistically. The absolute principal factor method was used to apportion the pollution sources of PM2.5 during the haze weather in Beijing city in winter 2013. The results showed that during the period of sampling, the volume concentration of Li, Mn, Pb, S etc. obeyed normal distribution approximately, and according to National Ambient Air Quality Standard issued by Ministry of Environmental Protection of the People’s Republic of China, the geometric mean concentration of As was twice the annual limit of standard reference, while Pb of some aerosol samples beyond the annual limit of standard reference respectively. The mass fraction of Fe, Zn, Pb, Ti accounted for over 0.1%, while that of Mn, Cu, As, Se etc. 0.01%. These elements were primary inorganic pollutants, and especially the hazards and sources of As and Pb should be concerned. There were 6 main pollution sources were chosen by the factor analysis method, including industrial dust and human beings activities, biomass combustion and building dust, soil and sand dusts, fossil fuel, electronic waste and metal smelting, with the variance contribution rate of 40.3%, 27.0%, 9.1%, 4.9%, 4.8% and 4.6% respectively. ICP-MS and ICP-AES can be applied to analyzing multi-elements in PM2.5 accurately and quickly to facilitate source apportionment, and it indicated that the relevant pollution sources should be considered and the effect of regional transferring of haze pollution sources should be taken into account, and specific measures should be taken for control.