Treatment outcome after surgical removal in oral carcinoma is poor due to inadequate methodologies available for marking surgical margins. Even though some methodologies for intraoperative margin assessment are under clinical and preclinical trials for other solid tumours, a promising modality for oral cancer surgery is not developed. Fluorescent-based optical imaging using Near Infrared (NIR) dyes tagged to tumour specific target will be an optimal tool for this purpose. One such target, Gastrin Releasing Peptide Receptor (GRPR) was selected for the study, and its binding peptide, TM1-IR680, was tested for its efficacy for surgical margin prediction in murine orthotopic model of oral cancer, derived from primary samples. Here, for the first time in a preclinical analysis, we show that the size and margin of oral cancer can be predicted, as revealed by 3D-imaging. Interestingly, the peptide was sensitive enough to detect lymph nodes that harboured dispersed tumour cells before colonization, which was impossible to identify by conventional histopathology. We recommend the use of TM1-NIR dyes alone or in combination with other technologies to improve the clinical outcome of oral cancer surgery.