The protein identification performs a crucial role in the contemporary medicine. Proteins may act as the potential biomarkers for investigating many diseases, e.g. the civilization-related ones.  Peptide mass fingerprinting (PMF) is a widely used protein identification method basing on mass spectrometry data. Economical reasons and time savings are of great importance in the identification experiments. Thereby, innovative ideas, which have the potential to improve the PMF identification, are still desired. A novel probability-based scoring scheme, which constitutes the last part of the PMF identification procedure, was developed. Presented scoring scheme incorporates an innovative idea, which assumes a different approach to modelling the distribution of proteins derived from the database, on the basis of which the score is computed. In the paper we assess a performance of the proposed scoring method against popular scoring scheme, i.e. Mascot (http://www.matrixscience.com/). The comparison of the methods includes scoring results obtained for the simulated data. Different levels of proteins samples contamination and different coverage of peptides sequences were considered in the empirical study.