Polygenic risk scores (PRSs) have been widely used to predict disease risk of complex traits. However, the predictability of PRSs has not reached to the level of clinical tools. It demands the development of more effective PRSs modelling technique. Previously, to evaluate the performance of PRSs, an independent validation dataset containing individual-level genotype is required. A recently developed method “SummaryAUC” makes it possible to evaluate PRSs model using only summary level statistics from an independent genome-wide association study (GWAS). With multiple candidate PRSs models, in practice, the one with the largest area under the ROC curve (AUC) is selected. It is often to have several PRSs models with similar values of AUC, but no statistical method is available to test their equivalence. In this project, we have developed a statistical method to test the equivalence of two AUCs using GWAS summary level data. Its performance is evaluated by simulation and real data analysis.