Aiming to a better understanding sources contributions and regional sources of fine particles, a total of 273 filter samples (159 of PM2.5 and 114 of PM1.0) were collected per 8 h during the winter 2016 at a southwest suburb of Beijing. Chemical compositions, including water soluble ions, organic carbon (OC), and elemental carbon (EC), as well as secondary organic carbon (SOC), were systematically analyzed and estimated. The total ions concentrations (TIC), OC, and SOC of PM2.5 were with the following order: 16:00–24:00 > 08:00–16:00 > 00:00–08:00. Since primary OC and EC were mainly attributed to the residential combustion in the night time, their valley values were observed in the daytime (08:00–16:00). However, the highest ratio value of SOC/OC was observed in the daytime. It is because that SOC is easily formed under sunshine and relatively high temperature in the daytime. Positive matrix factorization (PMF), clustering, and potential source contribution function (PSCF) were employed for apportioning sources contributions and speculating potential sources spatial distributions. The average concentrations of each species and the source contributions to each species were calculated based on the data of species concentrations with an 8 h period simulated by PMF model. Six likely sources, including secondary inorganic aerosols, coal combustion, industrial and traffic emissions, road dust, soil and construction dust, and biomass burning, were contributed to PM2.5 accounting for 29%, 21%, 17%, 16%, 9%, 8%, respectively. The results of cluster analysis indicated that most of air masses were transported from West and Northwest directions to the sampling location during the observation campaign. Several seriously polluted areas that might affect the air quality of Beijing by long-range transport were identified. Most of air masses were transported from Western and Northwestern China. According to the results of PSCF analysis, Western Shandong, Southern Hebei, Northern Henan, Western Inner Mongolia, Northern Shaanxi, and the whole Shanxi provinces should be the key areas of air pollution control in China. The exposure-response function was used to estimate the health impact associated with PM2.5 pollution. The population affected by PM2.5 during haze episodes reached 0.31 million, the premature death cases associated with PM2.5 reached 2032. These results provided important implication for making environmental policies to improve air quality in China.