Based　on　a　sample　of　Mergers　and　Acquisitions　of　China＇s　A-share　listed　companies　during　2003-2005,　this　paper　adopts　multivariate　adaptive　regression　splines　model　which　is　a　newly　developed　data　mining　technique　to　analyze　the　nonlinear　relationship　between　post-merger　operating　performance　and　M&A　characteristics　such　as　corporate　free　cash　flow,　holding　proportion　of　large　shareholders,　State-owned　share　ratio,　M&A　premium,　relative　target　size　and　industry　relatedness　of　target　and　merger.　The　result　demonstrates　that　corporate　free　cash　flow　per　share　has　a　nonlinear　threshold　effect　on　the　first-year　performance　of　merger　subsequent　to　the　M&A　transaction.　For　the　second　year　following　M&A　transaction,　corporate　free　cash　flow　per　share　significantly　negatively　impacts　the　performance.　However,　there　are　no　significant　evidence　that　the　M&A　characteristics　affect　the　performance　of　merger　for　the　third　year　after　M&A　transaction　and　the　year　of　M&A.　The　explanational　powers　of　our　regression　models　for　the　first　two　years　following　M&A　transaction　are　obviously　higher　than　that　of　previous　research,　which　suggests　that　our　regression　models　adequately　reflect　the　nonlinear　relationship　between　M&A　characteristics　and　post-merger　performance.　©　2010　IEEE.