Abstract: The responses of different scales to the phosphorus and potassium contents in leaves of Phyllostachys pubescens were studied. The results showed that It is of theoretical and practical significance to grasp the growth status and nutrient level of Phyllostachys pubescens in real time.Based on the original spectral spectrum of leaf and canopy of Phyllostachys pubescens as the data source, The spectral estimation models of phosphorus and potassium contents and sensitive bands at different scales were constructed by regression analysis using first order differential spectroscopy and so on. The stability and predictability were evaluated accurately. The results showed that the linear regression equation model with the first order differential spectrum of phosphorus content in the spectral analysis was 74.90%,And it has an advantage in establishing a model with higher specific phosphorus content. In the canopy spectrum analysis, the linear equation model with the first-order differential spectrum of potassium content is as high as 78.71%, and it is more advantageous in establishing the model with high specificity of potassium. At the same time, the first order differential spectrum can reduce the spatial heterogeneity to improve the spectral estimation model inversion accuracy.