The dynamic signature verification has problems of the unequal length of dynamic features, complex dynamic signature verification methods, and low recognition rate. To address these issues, a dynamic signature verification method based on correlation coefficient is proposed. First, the feature weight sum in the corresponding region is filtered and calculated by dividing the original feature region. And the correlation coefficients between signature features are calculated by the Pearson correlation analysis method. Secondly, the Pearson correlation coefficient distribution of genuine and simulated signatures is analyzed with the correlation coefficient as a new feature. Finally, the signature is evaluated by combining the Gaussian density function model and setting an individual discrimination threshold. Experimental results show that the Pearson correlation coefficient inside the genuine signature is generally higher than that between the genuine and simulated signatures. This method shows better signature verification performance on SVC and xLongSignDB data sets. The false rejection rate and false acceptance rate on xLongSignDB data sets are 2. 1% and 1. 7% , respectively.