Abstract: Sifting a target patent with core competitiveness through existing patents is the key step of patent circumvention to raise the starting point. For massive patents in design domain, a multi-index evaluation model was proposed, which effectively integrates subjective and objective weights. To improve the scientificalness and operability of evaluation indicators, the positive correlations between patent bibliographic data and patent quality were analyzed. As a result, five bibliographical data, i.e. patent duration, claims number, family number, citing number and cited number, were selected as indexes to establish a comprehensive evaluation system of patent quality. By combining Delphi Method, Analytic Hierarchy Process and Mean Square Error Decision Method, the objective and subjective weighting approach was presented to ensure authority of the evaluation result and reduce effects of personal bias and compromises. A process model of selecting target patents based on the patent quality evaluation model was built and applied in a real case of patent circumvention. The results show that the model improves the overall efficiency of finding target patents with core competitiveness.