Abstract: Based on the hourly monitoring data of near-surface PM_2.5 mass concentration and the meteorological data of the same period in Kunming area from 2016 to 2020, the correlation between the variation characteristics of PM_2.5 mass concentration and meteorological factors was analyzed, and the seasonal characteristics of PM_2.5 transport path and pollution track and the distribution of potential source areas were studied by using the backward trajectory model. The results showed that the mass concentration of PM_2.5 increased first and then decreased and fluctuated peak-valley throughout the year. The overall concentration was 57.4% higher in spring and winter than in autumn and summer, and relatively stable at night, while the variation range was large in daytime, and the concentration at night was significantly higher than that in daytime (22.6%). The mass concentration of PM_2.5 was negatively correlated with air temperature and wind speed, the correlation coefficients were -0.60 and -0.13, respectively, and was positively correlated with air pressure (R=0.44, P < 0.01). There were seasonal differences in the source direction of airflow trajectories in Kunming area. The sources of spring and winter trajectories were concentrated, mainly from the southwest, and scattered in summer and autumn centered on the south. Except for all the airflows arriving in Kunming area in summer and autumn are clean airflow, the transmission path in spring is far and the contribution source area is large. The central part of Yunnan province (Yuxi city), the southwest Yunnan province (Chuxiong prefecture, Puer city) and parts of northwestern Bangladesh and northern Myanmar are the most important contributors to PM_2.5 quality concentrations in Kunming area, and the coverage and contribution degree of high values in winter are smaller than those in spring, and are only distributed in dots in south-central Yunnan province (Yuxi city, Honghe prefecture).