We present a new approach for peptide identification from a list of candidates using collision-induced dissociation tandem mass spectrometry data. This approach is based on a probabilistic model for the occurrence of spectral peaks corresponding to frequently observed partial peptide ions. As part of the identification procedure, a probability score is produced that indicates the likelihood of any given candidate being the correct match. The performance of the proposed algorithm is tested on a large test dataset consisting of spectra from charge = 2+ tryptic peptides of different lengths ranging from 6-mers to 30-mers. Initial results suggest that this approach works well with error rates typically below 5%.