This　paper　presents　a　collaborative　filtering　recommendation　system　combined　with　market　value　functions.　In　the　approach,　by　building　market　value　function　model　for　each　item,　we　can　produce　a　user　rating　information　table.　Then　Collaborative　Filtering　algorithm　can　produce　a　list　of　recommended　items.　Experiment　in　real　world　examples　shows　that　the　method　is　effective.