This study evaluated breast cancer cells and metastasis potential using visible (532 nm) resonance Raman spectroscopy (VRRS). Three cell lines were investigated, MCF-10 (fibrocystic disease (benign)), MCF-7 (estrogen- and progesterone-receptive invasive ductal carcinoma (IDC)), and MDA-MB-231 (triple negative IDC). Peak analysis and multivariate unmixing methods including principal component analysis, partial least squares and nonnegative matrix factorization along with support vector machines were used to classify the samples. The cell lines were accurately classified with leave-one-out cross-validation. Optimal features were selected using a wrapper feature selection algorithm which helped to identify key biomarkers related to metastasis potential of the cell lines.