A　part-time　hybrid　energy　management　strategy　based　on　energy　prediction　is　proposed　for　range-extended　electric　vehicle　in　this　paper.　Firstly　according　to　static　navigation　data,　two　energy　prediction　algorithms　based　on　moving　average　and　emergency　respectively　are　devised　by　utilizing　the　principle　of　decision　tree　algorithm.　Then　the　two　prediction　algorithms　are　tested　with　their　features　analyzed　respectively.　Finally,　according　to　historical　and　future　data　simulated,　the　situation　of　energy　use　is　predicted,　the　accuracy　of　prediction　is　analyzed,　and　the　energy　distributions　of　part-time　hybrid　energy　management　strategy　with　two　different　prediction　algorithms　are　compared.　The　results　show　that　moving　average-based　prediction　is　better　than　emergency-based　prediction　no　matter　for　the　gradual　increase　of　SOC　in　initial　stage,　its　mutation　in　middle　stage　or　its　fluctuation　in　end　stage.　©　2017,　Society　of　Automotive　Engineers　of　China.　All　right　reserved.