α-Bisabolol is a commercially important aroma chemical currently obtained from the Candeia tree (Vanillosmopsiserythropappa). Continuous unsustainable harvesting of the Candeia tree has prompted the urgent need to identify alternative crops as a source of this commercially important sesquiterpene alcohol. A chemotaxonomic assessment of two Salvia species indigenous to South Africa is presented and recommended as a potential source of α-bisabolol. The essential oil obtained by hydrodistillation of the aerial parts was analysed by gas chromatography coupled to mass spectrometry (GC–MS) and mid-infrared spectroscopy (MIRS). Orthogonal projections to latent structures-discriminant analysis (OPLS–DA) were used for multivariate classification of the oils based on GC–MS and MIRS data. Partial least squares (PLS) calibration models were developed on the MIRS data for the quantification of α-bisabolol using GC–MS as the reference method. A clear distinction between Salvia stenophylla and S. runcinata oils was observed using OPLS–DA on both GC–MS and MIRS data. The MIR calibration models showed high coefficients of determination (R2= 0.999) and low errors of prediction (RMSEP = 0.540%) for α-bisabolol content.