In this study, the nitrogen content of rice plants were measured by remote sensing using multispectral
imaging techniques. The multi-spectral imaging model using 4 band wavelengths at 555, 660, 680
and 780nm was successfully developed to predict the nitrogen content in rice (Oryza sativa L. cv. Tainung
67). The model gave results of r =0.84, SEC = 0.21, RSEC=12.7% for calibration and r = 0.77, SEP = 0.31,
RSEP = 15.5% for prediction. This model was applied to predict nitrogen content in different variety of rice
(Taiken 14) and gave results of r=0.72, SEP=0.29, RSEP=17.3%. The prediction results of Tainung 67 in
different growth stage and year gave the results of r=0.70, SEP=0.35, RSEP=17.3%. The model exhibited
the feasibility of wide applications to different rice varieties and growth years. The model was further applied
to the analysis of the distribution of nitrogen content in rice plants in a large scale of paddy field by taking
multi-images at 15m above the ground using a mobile lift. The information of crop’s nitrogen distribution
could contribute to the development of a decision support system for a site-specific crop management in
precision agriculture.