Raman spectroscopy, with its ability to detect substances at the molecular level and its non-labeling, high-precision, and non-invasive characteristics, is widely applied within the biomedical field. Compared with non-biological samples, biological samples exhibit a higher complexity in terms of material composition and distribution, coupled with substantial individual variances. Moreover, the Raman scattering signals emitted by these biological samples are inherently weak. Therefore, difficulties such as faint signals, data collection obstacles, and analysis complexities are often encountered during their Raman detection. With the advancement of computer technology, machine learning has played a significant role in various fields such as biomedicine, optics, and chemistry in recent years. It has performed well in tasks like classification, data analysis, and data prediction. In response to the challenges associated with Raman spectroscopy examination of biological samples, many researchers in recent years have employed a combination of Raman spectroscopy and machine learning for the analysis of biological samples at various scales, which has realizesd functions such as cancer diagnosis, cell sorting, bacteria identification, and virus monitoring. The combination of Raman detection and machine learning thus opens up a new research avenue in the realm of biomedicine.