Introduction To perform large scale metabolomic analyses, high throughput approaches are required. The direct introduction mass spectrometry (DIMS) approach appears to be very attractive to achieve this goal. However, processing DIMS data is still very challenging due to the large number of samples and the intrinsic complexity of the mass spectra. Objectives The objective of this study is to develop a computational procedure, based on an innovative chemometric method, i.e. Independent component–discriminant analysis (IC–DA), for processing DIMS data. Method Metabolomic fingerprints were obtained by direct introduction high resolution mass spectrometry (DI-HRMS) analysis of urine samples of subjects that had been professionally exposed to pesticides. Spectral data were processed using the developed IC–DA procedure. Results obtained from this method were compared to those obtained by the conventional Partial least squares–discriminant analysis (PLS–DA). For both the IC–DA and PLS–DA methods, a validation was performed based on a permutation test. Result IC–DA results enabled a good detection of discriminant variables and a clear discrimination of control samples and exposure classes whereas a less striking discrimination was obtained with PLS–DA. Putative annotation of these variables was performed using metabolomic databases. Targeted correlation analysis was used for the detection of ions associated with the most discriminant variables, consolidating their identity assignment. Conclusion This study demonstrated the efficiency of IC–DA to discriminate the different exposure groups. As well the improvement of high throughput metabolomic studies was provided by combining DI–HRMS with this new chemometric tool.