Extracting biochemical information from cell and tissue autofluorescence without using any external biomarkers is a promising approach for disease diagnosis and treatment. This thesis reports the advancement of fluorescence microscopy and image acquisition tools to observe autofluorescence signals, and I intended to use these signals as a biomarker for disease diagnosis and classification. In this research, I applied the hyperspectral imaging technique in observing the autofluorescence signals in peripheral blood monocytes from both healthy control and motor neuron disease patients. Current motor neuron disease diagnosis and classification methods are mainly based on clinical observations, which depends upon an experienced neurologist to make an informed clinical judgment. The development of an early-stage biomarker would be beneficial for the diagnosis of this disease. Research presented in this thesis centers around the identification of distinct autofluorescence features from the peripheral blood monocytes, which can help reveal the difference between a healthy person and a motor neuron disease patient. From my result, a novel biomarker is discovered by applying the hyperspectral image technique. It is based on the mean cell intensity per unit area on a selection of channels. According to 15 samples, the classification methods showed promising results. With further validation, it may be applied in motor neuron disease classification.