With the rapid development of urbanization, air pollution, especially fine particulate matter (PM_2.5), has become an important factor restricting urban environment. Relevant studies have shown that road traffic is one of the important sources of PM_2.5 among many sources, and plant communities in road greenbelt could reduce PM_2.5. In this study, a combination of site observation and ENVI-met simulation was used to investigate the influence of plant communities in urban road greenbelt on PM_2.5. We also analyzed the influence of plant community configuration on PM_2.5 in three modes including tree (i), tree+hedge (ii), and tree+hedge+shrub (iii), so as to reveal the influence of length, width, height and LAI on PM_2.5. The results showed that (1) in field observation, the reduction rate of PM_2.5 in the three configuration modes of plant community of road greenbelt was iii (14.2%) >ii (12.9%) > i (9.2%). (2) The increase of plant species, greenbelt width, and leaf area index played a positive role in PM_2.5 reduction, while the increase of height and length played a negative role. (3) A certain amount of reduction in the length of the green space formed a region behind the green space that had a lower PM_2.5 concentration than the longer length environment. The width of this area was about 80% of the length of the green space and decreased as the distance from the green space increased. However, the PM_2.5 concentration of other spaces without green space would increase. (4) The increase in green space height would force PM_2.5 to higher space.