The present study aims to identify the expression of transgene in given rice plant samples in certain conditions. To avoid external noise caused by temperature change and water-loss,field spectrum was collected with ASD field spectrometer in natural state. The study calculated the mean spectrum of samples as main data set analyzed which were controlled by inner clustering coefficient to ensure data quality. By mean spectrum,the noise from random distinctions in few individual cultivators,which could not be expressed in the class stably,could be weakened even with filtering. With the help of parameters,such as red edge and green peak,this study gave qualitative spectral differences between transgenic samples and their parents. The results show that the transgenes in rice plant were expressed and influenced the samples. Moreover,it was found that the parameters of area are more suitable for describing the differences/changes of the samples,while PRI and SR-PRI are more sensitive to indicate them. Most of the above results could be found on the continuum-removal spectrum curve of samples. These conclusive results demonstrate that hyperspectral remote sensing technique has good prospects and application potential in transgene expression detection and monitoring,especially in plant breeding process.