A novel aspect of principal component analysis (PCA) which was recently revealed has been applied to the separation of a Raman signal from a strong fluorescence emission background. A Raman spectrum of indene with weak intensity that was almost hidden in the high background was readily resolved by PCA only. The resolved spectrum closely resembled the pure Raman spectrum without disturbance by the background. Another resolved spectrum by use of target testing was also very similar to the resolved spectrum by PCA alone. These results preliminarily suggest that PCA is useful to obtain Raman spectra from a collection of replicated raw spectra that are dominated by strong background signal.