<p>Diagnosis and determining the surgical margin of breast cancer are challenging tasks in oncology, so there is a need for an effective and efficient diagnostic process. THz time-domain spectroscopy (THz-TDS) has been proposed as a promising non-invasive technique for tissue characterization due to its sensitivity to water and molecular content. In this paper, these dielectric properties (refractive index, absorption and permittivity) as well the white tissue (fibrotic tissue) and yellow tissue (adipose tissues) have analyzed to show the potential of THz-TDS at our frequency range in different phase identification. Experimental results showed that white breast tissue possessed higher refractive index (1.8 at 0.4 THz) and permittivity (3.24 at 0.4 THz) than those of yellow tissue (1.4 and 2.56), which originated from the disparity in water content and structural organization in each type of the tissues. Moreover, the absorption coefficient of white tissue was much larger than that of yellow tissue, which meant that the interaction between THz waves and white tissue was stronger. The observed decrease in permittivity with an increase of frequency was indicative of less interaction between THz waves at higher frequencies. These results indicate the capability of THz-TDS for immediate real-time label-free quantitative discrimination of breast tissues, which can potentially improve margin evaluation and breast cancer diagnosis. However, several issues need to be addressed further, including but not limited to tissue desiccation in ex vivo settings, inter-sample variability, and pathological effects. Subsequent studies should also further in vivo validation and the incorporation of machine learning in order to categorize tissues automatically and improve diagnostic accuracy. This work further validates the applicability of THz-TDS in biomedical spectroscopy and strongly supports its potential as a novel device for breast cancer diagnostics and tissue assessment.</p>