In order to estimate the nitrogen content of sugar beet leaves quickly, sugar beet was taken as the research object. The hyperspectral image data of canopy leaves was obtained by hyperspectral imaging spectrometer. At the same time, the nitrogen content of leaves was determined by Kjeldahl method. Based on the meticulous sampling method, the normalized spectral parameter (NDSI) and the soiladjusted vegetation index (SASI) were constructed in the wholewavelength range. Moreover, in order to search for the optimal value of L in SASI under arbitrary band combination，the particle swarm optimization algorithm was proposed to optimize the L. On the basis of the previous work mentioned above, the sensitive spectral parameters were selected to achieve the optimization, and the estimation model was constructed to carry out the quantitative diagnosis and visualization research of the nitrogen content in the sugar beet leaves. The results indicated that the sensitivity of SASI to the canopy leaf nitrogen content (CLNC) of sugar beet was higher than that of NDSI for each different growth period. Especially in the nearinfrared region where saturation easilyoccurred, the correlation was significantly improved. Compared with the conventional spectral parameters, based on SASI1(R430.20, R896.76) and SASI2(R433.03, R896.01), an optimal CLNC estimation model of BP net for the rapid growth period of the beet leaves was able to be established. The determination coefficient (R2) of validation set was 0.78, the root mean square error (RMSE) was 2.48g/kg and the relative error (RE) was 4.18% (in the year of 2015). The model established based on SASI3(R952.09, R946.11) and SASI4(R760.37, R803.48) for the sugar growth period had the best performance. The R2 of verification set was 0.67, the RMSE was 2.71g/kg, and RE was 4.72% (in the year of 2015). The optimal modeling parameters for the sugar accumulation period were SASI5 (R883.30， R887.79), and the R2 of the model was 0.72, the RMSE was 2.54g/kg, and the RE was 4.49% (in the year of 2015). Based on the above model, combined with the spectral information of each band under every pixel of hyperspectral image, the CLNC was calculated, and the CLNC concentration graphs of sugar beet were plotted, which directly and visually presented the distribution of nitrogen content in the sugar beet leaves at different time scales and different leaf positions. The research results introduced that the proposed estimation method of CLNC in sugar beet was feasible, which also provided technical support for timely observation of crop growth and nutritional diagnosis.