Due　to　the　correlation　and　diversity　of　robotic　kinematic　dexterity　indexes,　the　principal　component　analysis　(PCA)　and　kernel　principal　component　analysis　(KPCA)　based　on　linear　dimension　reduction　and　nonlinear　dimension　reduction　principle　could　be　respectively　introduced　into　comprehensive　kinematic　dexterity　performance　evaluation　of　space　3R　robot　of　different　tasks.　By　comparing　different　dimension　reduction　effects,　the　KPCA　method　could　deal　more　effectively　with　the　nonlinear　relationship　among　different　single　kinematic　dexterity　indexes,　and　its　calculation　result　is　more　reasonable　for　containing　more　comprehensive　information.　KPCA＇s　calculation　provides　scientific　basis　for　optimum　order　of　robotic　tasks,　and　furthermore　a　new　optimization　method　for　robotic　task　selection　is　proposed　based　on　various　performance　indexes.　©　Copy　Right　by　Hight　Technology　Letters　Press.