The　traditional　valuation　methods　developed　in　industrial　economy　can＇t　accurately　appraise　firms　in　knowledge　economy　due　to　their　ignorance　of　intellectual　capital.　Based　on　the　value　creating　mechanism　of　intellectual　capital,　this　study　proposes　a　novel　valuation　approach　named　hierarchical　clustering-MARS　approach.　The　method　intelligently　searches　similar　firms　as　target　firm　through　hierarchical　clustering,　and　utilizes　nonparametric　multivariate　adaptive　regression　splines　(MARS)　technique　to　estimate　the　value　creation　function　for　target　firm.　Then　the　target　firm　value　is　forecasted　using　the　estimated　value　creation　function.　The　study　also　investigates　the　effectiveness　of　the　new　approach　by　a　sample　of　Chinese　high　technology　firms　listed　in　A-share　market.　The　empirical　result　proves　that　the　hierarchical　clustering-MARS　approach　meets　the　accuracy　requirement　of　appraisal　and　is　of　practical　use.　©　2011　IEEE.