More researches were carried out to investigate the application of spectral technology on crop nutrition diagnosis. However, the complex conditions in the field results in the uncertainty of spectrum. In this paper, the influence of foliar dust on spectral of the crop beside the nation road was studied, the differences of the raw spectral reflectance and first derivative spectral reflectance between the foliar with dust and the foliar washed with deionised water were analyzed, nitrogen prediction models were built on the disturbance of foliar dust. Result showed that the dust foliar spectral reflectance increased in the visible light (350～720 nm) and shortwave infrared (SWIR) (1 360～2 500 nm) regions whereas the spectral reflectance in the near infrared (NIR) (720～1 360 nm) wavelength regions decreased. There were no change rules for blue edge position, yellow edge position, red edge position, blue edge slope and yellow edge slope on the effect of foliar dust, but red edge slope, blue edge area, yellow edge area, red edge area decreased. Determinate coefficient(R2) of nitrogen prediction models reduced in the condition of foliar dust. The primary research work about the condition of foliar dust was studied which helps to provide foundation for evaluating effect of foliar dust and proposing foliar dust modification model in the future.