The identification of the most competent embryos for transfer to the uterus constitutes the main challenge of in-vitro fertilization (IVF). We established a metabolomic-based approach applying Fourier Transform Infrared spectroscopy (FTIR) on 130 samples of 3-days embryo culture supernatants from 26 embryos that implanted and 104 that failed. Examining the internal structure of the data by unsupervised multivariate analysis, it was observed that the supernatants of nonimplanted embryos contained highly heterogeneous spectral features. These features were overlapping with metabolic-implantation fingerprints, thus demonstrating that in establishing embryo-assessment models a one-class modelling involving only the samples with positive-implantation outcomes should be applied. Analysis of variance confirmed that the women´s age (>40 years) undermined the implantation of the embryos exhibiting implantation metabolomics, and also that constituted a condition triggering embryos to express nonimplantation metabolomics. We conclude that IVF-success rates can be significantly improved if FTIR spectroscopy is used as an embryo-selection criterion.