Urban green plants provide purification functions for the urban ecosystem, which play a variety of environmental protection roles such as air purification, dust retention and dust fall. On the other hand, dust retention can also affect green plants in turn. To study how dust retention affects the spectral characteristics of urban plant leaves, the point cloud data of four kinds of evergreen leaf samples (Fatsia japonica, Photinia stenophylla, Deyeuxia langsdorffii and Magnolia grandiflora) are collected with hyperspectral lidar, and then the effect of dust retention on leaf spectral characteristics is analyzed. The results show that dust retention has an influence on the reflectance in visible light band despite of plant species, and has a great impact on the reflectance difference in the near-infrared band for the same kind of leaves. The reflectance difference in visible light band is 1.21%～3.41%, and that in near infrared is 1.76%～8.49s of leaves to dust retention is the smallest (less than 3.7%), that of the ratio vegetation index (RVI) is the largest (more than 20.0%, except Deyeuxia langsdorffii), and that of the red edge index (SDr), simple ratio index (SR) and leaf chlorophyll index (LCI) is low and unstable for all leaves. Furthermore, the fitting models based on the correlation between the dust retention vegetation index and responsiveness are established and tested, it is shown that the best one is based on LCI, which can be expressed as b = -1.527a + 0.659, and the determination coefficient R2 is about 0.88.