High-resolution mass spectrometry coupled with pattern recognition techniques is an established tool to perform comprehensive metabolite profiling of biological datasets. This paves the way for new, powerful and innovative diagnostic approaches in the post-genomic era and molecular medicine. However, interpreting untargeted metabolomic data requires robust, reproducible and reliable analytical methods to translate results into biologically relevant and actionable knowledge. The analyses of biological samples were developed based on ultra-high performance liquid chromatography (UHPLC) coupled to ion mobility - mass spectrometry (IM-MS). A strategy for optimizing the analytical conditions for untargeted UHPLC-IM-MS methods is proposed using an experimental design approach. Optimization experiments were conducted through a screening process designed to identify the factors that have significant effects on the elected responses (total number of peaks and number of reliable peaks). For this purpose, full and fractional factorial designs were used while partial least squares regression was used for experimental design modeling and optimization of parameter values. The total number of peaks yielded the best predictive model and is used for optimization of parameters setting.