This　paper　proposes　a　novel　neural　dynamical　approach　to　a　class　of　convex　quadratic　programming　problems　where　the　number　of　variables　is　larger　than　the　number　of　equality　constraints.　The　proposed　continuous-time　and　proposed　discrete-time　neural　dynamical　approach　are　guaranteed　to　be　globally　convergent　to　an　optimal　solution.　Moreover,　the　number　of　its　neurons　is　equal　to　the　number　of　equality　constraints.　In　contrast,　the　number　of　neurons　in　existing　neural　dynamical　methods　is　at　least　the　number　of　the　variables.　Therefore,　the　proposed　neural　dynamical　approach　has　a　low　computational　complexity.　Compared　with　conventional　numerical　optimization　methods,　the　proposed　discrete-time　neural　dynamical　approach　reduces　multiplication　operation　per　iteration　and　has　a　large　computational　step　length.　Computational　examples　and　two　efficient　applications　to　signal　processing　and　robot　control　further　confirm　the　good　performance　of　the　proposed　approach.　©　2009　Elsevier　Ltd.　All　rights　reserved.