Multi-node　sensors　for　human　posture　detection,　by　acquiring　kinematic　parameters　of　the　human　body,　helps　to　further　study　the　laws　of　human　motion.　It　can　be　used　as　reference　for　the　quantitative　analysis　tool　for　some　specific　application,　such　as　healthcare,　sports　training　and　military　affairs,　etc.　Compared　with　the　traditional　optical　method,　posture　detection　based　on　the　inertial　sensors　has　smaller　limitation　of　space,　lower　cost,　and　easier　implementation.　In　this　paper,　a　human　posture　detection　system　was　introduced.　Utilizing　the　parameter　data　obtained　by　the　inertial　sensors,　three-dimensional　angles　of　the　human　hand　movement　could　be　calculated　via　quaternion　algorithm　for　data　fusion.　The　angles　data　transmission　among　the　sensor　nodes　was　successfully　realized　by　the　human　body　communication(HBC)　transceivers　based　on　capacitive　coupling　of　multi-carriers　OOK　modulation　at　the　data　rate　of　57.6kbps.　The　bit　error　rate(BER)　was　less　than　10-5.　The　human　posture　could　be　reconstructed　on　the　PC　host.　Ultimately,　the　implementation　of　the　overall　result　showed　the　feasibility　of　the　system.　©　2016　Croatian　Society　MIPRO.