To completely understand the influence of temperature and relative humidity in the boundary layer on the vertical distribution of PM2.5 and near-surface pollution, in this study, an unmanned aerial vehicle (UVA) equipped with multiparameter atmospheric environment detector was used for the vertical observation of PM2.5 concentration, temperature, and relative humidity in Nanjing from 3 December to 4 December 2017 and from 23 December to 24 December 2017. Combined with the analysis of meteorological data and the application of HYSPLIT4 (Hybrid Single Particle Lagrangian Integrated Trajectory version 4) trajectory calculation model, the vertical distribution characteristics of PM2.5 and the causes of PM2.5 pollution processes from 3 December to 4 December 2017 and from 23 December to 24 December 2017 were analyzed. The results showed a significant positive correlation between PM2.5 concentration and relative humidity. Moreover, the average correlation coefficient reached 0.96 in the six observations from 23 December to 24 December , 2017. Under the inversion layer, the PM2.5 concentration and relative humidity were high and the vertical difference was small. Over the inversion layer, the PM2.5 concentration and relative humidity rapidly decreased with the increase in height. Because of the poor atmospheric diffusion conditions, PM2.5 continuously accumulates in the south of the North China Plain and moves to the south and southeast under the influence of the high-pressure system. Both PM2.5 pollution processes were significantly affected by external transport. Atmospheric inversion significantly inhibited the upward transport of PM2.5 and water vapor. Moreover, external transport and local inversion were the main causes of these two PM2.5 pollution processes.