The　stochastic　driving　cycles　pose　a　challenge　to　the　actual　implementation　of　the　control　strategy　for　fuel　cell　electric　vehicles　(FCEVs).　Considering　the　problem,　this　research　proposes　a　predictive　control　strategy　based　on　velocity　prediction.　Firstly,　a　novel　velocity　prediction　method　is　developed,　which　considers　the　prediction　error　of　back　propagation　neural　network　(BPNN)-based　method.　Then,　it　is　incorporated　into　the　predictive　control　strategy　for　a　plug-in　FCEV.　Finally,　simulation　studies　are　conducted　to　verify　the　effectiveness　and　superiority　of　the　proposed　predictive　control　strategy.　Simulation　results　show　that　the　proposed　velocity　prediction　method　can　adaptive　to　different　driving　cycles　with　high　accuracy.　In　another　case,　with　the　velocity　prediction　used　in　the　predictive　control　strategy,　hydrogen　consumption　reduces　by　17.07%　when　compared　with　the　traditional　rule-based　strategy.　All　the　results　indicate　that　the　designed　velocity　prediction　approach　made　it　possible　to　forecast　vehicle　velocity　with　relatively　high　precision　and　promising　to　promote　the　energy　management　strategy　(EMS)　to　reduce　the　hydrogen　consumption　of　the　plug-in　FCEV.