Abstract: There are a large number of performance shaping factor (PSF) summed up from typical human reliability analysis (HRA) methods. PSF is generally judged by experts in order to quantify the basic human error probability (HEP) in HRA. Hence it is subjective, vague and uncertain. An evaluation model of PSF was constructed based on the combination of correlation coefficient matrix, distance classification of graphs and principal component analysis (PCA) in the digital main control room (MCR) of nuclear power plant (NPP). The purpose was to identify the main PSF which affects human performance in different types of human error events for making decisions to reduce human error. 179 post-event reports of one NPP were provided to illustrate the evaluation, and the results indicate that the evaluation model can effectively evaluate the PSF of the operators in the digital MCR of NPP.