Glioblastoma (GBM) is the most common and aggressive malignant brain tumour in adults. Patient survival rates are strongly dependent on the successfully resection of the tumour. In this framework, multimodal optical spectroscopy could provide a fast and label-free tool for improving tumour detection and guiding the removal of diseased tissue. In this study, we used an optical fibre-probe system combining multiple spectroscopic techniques for in vivo examination of normal and GBM tissues in mouse brain. Specifically, the probe – based on a fibre-bundle with optical fibres of various size and properties – allowed performing spectroscopic measurements based on fluorescence, Raman, and diffuse reflectance spectroscopy though two optical windows implanted on the head of each animal. Two visible laser diodes were used for fluorescence spectroscopy, a laser diode emitting in the NIR was used for Raman spectroscopy, and a fibre-coupled halogen lamp for diffuse reflectance. All spectral recordings were done when the animals were anesthetized; optical inspection required less than 4 minutes for each animal. The recorded data were analysed using Principal Component Analysis (PCA) for obtaining an automated classification of the examined tissues based on the intrinsic spectral information provided by Raman and reflectance spectroscopy. The presented method demonstrated high sensitivity and specificity in discriminating GBM from normal brain. Furthermore, we found that the multimodal approach is crucial for improving diagnostic capabilities beyond what can be achieved from individual techniques.