The overexpression of HER2 (human epidermal growth factor receptor 2) in breast cancer is associated with increased disease recurrence and worse prognosis. Current diagnosis of HER2 positive breast cancer is time consuming with an estimated 20% inaccuracy. Raman spectroscopy is a proven method for pathological diagnosis based on the molecular composition of tissues. This study aimed to determine the feasibility of Raman spectroscopy to differentially identify the amplification of HER2 in cells. Three cell lines including BT474 (HER2 overexpressing breast cancer cell), MCF-10A (human breast epithelial cell), and MCF-10A with overexpressing HER2, were investigated using a bench top confocal Raman system. A diagnostic algorithm based on generalized linear model (GLM) with elastic-net penalties was established to discriminate 318 spectra collected from the cells, and to identify the spectra regions that differentiate the cell lines. The algorithm was able to differentially identify BT474 breast cancer cells with an overall sensitivity of 100% and specificity of 99%. The results demonstrate the capability of Raman spectroscopy to determine HER2 status in cells. Raman spectroscopy shows promise for application in the diagnosis of HER2 positive breast cancer in clinical practice.