Nowadays development of industry and traffic are the main contributor to city air pollution in the city of GuangZhou, China. Conventional methods for investigating atmosphere potentially harmful element pollution based on sampling and chemical analysis are time and labor consuming and relatively expensive. Reflectance spectroscopy within the visible-near-infrared region of vegetation in city has been widely used to predict atmosphere constituents due to its rapidity, convenience and accuracy. The objective of this study was to examine the possibility of using leaves reflectance spectra of vegetation as a rapid method to simultaneously assess pollutant (S, Cd, Cu, Hg, Pb, XCl, XF) in the atmosphere of the Guangzhou area. This article has studied the spectral features of polluted leaf surface of Ficus microcarpa in 1985 and 1998. According to the analysis, comprehensive assessment for the change of atmospheric condition and degrees of pollution were given. This conclusion was confirmed by the monitored data got from chemical analysis. Future study with real remote sensing data and field measurements were strongly recommended.