In this study data generated by H-1 NMR were combined with chemometrics to analyse brain and plasma samples from APP/PS1 and wild type mice with the aim of developing a statistical model capable of predicting the features of Alzheimer's disease (AD) displayed by this animal model. APP/PS1 is a well characterised double transgenic mouse model of AD and the results here demonstrate the potential of NMR technology as a platform for the detecting this disease. Using partial least squares discriminant analysis a model was built using both brain extracts (R-2 = 0.99; Q(2) = 0.66) and a high throughput method of plasma analysis (R-2 = 0.98; Q(2) = 0.75) capable of predicting AD in APP/PS1 mice. Analysis of brain extracts led to the elucidation of 20 metabolites and 16 of these were quantifiable. Relative brain levels of ascorbate, creatine, gamma-aminobutyric acid and N-acetyl aspartic acid were significantly altered in APP/PS1 mice (p <0.05). Analysis of plasma identified 14 metabolites and the levels of acetate, citrate, glutamate, glutamine, methionine, and an unknown signal were significantly altered in APP/PS1 mice (p <0.05). Combining H-1 NMR spectral data with chemometrics has been previously used to study biochemical disturbances in various disease states. This study further indicates the translational potential of this technology for identifying AD in people attending the memory clinic.