The aim of this study is to develop a novel technique for improving the intraoperative margin assessment of glioblastoma by examining the total extrinsic extracellular matrix (ECM) with eosin staining using fluorescence lifetime imaging microscopy (FLIM) and scale-invariant feature transform (SIFT) descriptor analysis. Pseudo-color FLIM images obviously exhibit ECM distributions, changes in sequential sections, and different regions of interest. Meanwhile, SIFT descriptors are first utilized for the discrimination of glioblastoma margins by matching similar ECM regions and extracting keypoint orientations from FLIM images obtained from a series of continuous slices. The findings indicate that FLIM imaging with SIFT analysis of the total ECM is a promising method for improving intraoperative diagnosis of frozen and surgically excised brain specimen sections.