Oral cavity cancers are the fourth common cancers in companion dogs. Common malignant oral tumors in pet dogs include squamous cell carcinoma, melanoma and fibrosarcoma while benign tumors include epulids, papilomas and extramedullary plasmacytomas. Malignant tumors are often aggressive with potential to invade locally and distantly. The occurrence of malignant and benign lesions are evenly matched. Hence early diagnosis is important for management. Current diagnosis is based on aspiration cytology and biopsy, which is a time consuming and invasive procedure. Hence less invasive, reliable and quicker diagnostic technique is preferable. Raman spectroscopy-based studies have demonstrated the feasibility of classifying normal, premalignant and malignant oral conditions in human subjects as well as rodent animal models. The present study was undertaken to explore feasibility of classifying canine tumour and normal oral tissues with Raman spectroscopy (RS). RS is a non-destructive vibrational spectroscopic technique, which provides information about molecular composition, molecular structures and molecular interactions in tissue. Raman spectra of histopathologically confirmed 13 (10-abnormal and 3- normal) samples were acquired using a Raman Confocal Microscope (Raman alpha300R, WITEC, GMBH, GERMANY). Spectral acquisition parameters were : laser power-28mW, integration-10 s and averaged over 10 accumulations. Spectra were pre-processed and subjected to unsupervised Principal-Component Analysis (PCA) to identify trends of classification. Supervised LDA (Linear-Discriminant Analysis) showed high classification efficiency. Findings of this explorative study suggest that Raman spectroscopy can be developed as a non-invasive, label free, early diagnostic tool for cancers in pet dog.