A multi-parametric sequence-specific model for predicting peptide electrophoretic mobility has been developed using large-scale bottom-up proteomic CE-MS data (5% acetic acid as background electrolyte). Peptide charge (Z) and size (molecular weight, M) are the two major factors determining electrophoretic mobility-- in complete agreement with previous studies. The extended size of the dataset (>4000 peptides) permits access to many sequence-specific factors that impact peptide mobility. The presence of acidic residues Asp and Glu near the peptide N-terminus is by far the most the prominent among them. The induction effect of the side chain of N-terminal Asp reduces the basicity of the N-terminal amino group- and as hence its charge- by ~0.27 units, lowering mobility. The correlation of the model (R2~0.995) indicates that the peptide separation process in CZE is relatively simple and can be predicted to a much higher precision than current RP-HPLC models. Similar to RP-HPLC prediction studies, we anticipate future studies that introduce peptide migration standards, collect larger datasets for modeling through the alignment of multiple CZE-MS acquisitions, and study of the behaviour of peptides carrying post-translational modifications. The increased size of datasets will also permit investigation of the fine-scale effects of peptide secondary structure on peptide mobility. We observed that peptides with higher helical propensity tend to have higher than predicted electrophoretic mobility; the incorporation of these features into CZE migration models will require significantly larger data sets.